A direct test of whether feedback produced a fix
GAO anonymously reported accessibility problems to 23 federal agencies. Fourteen corrected the issues between September 2025 and February 2026; nine had not corrected them by May.
That produces a 60.9% repair rate and a 39.1% unresolved rate in the tested group. The test does not measure every page or system at those agencies, but it answers a narrower accountability question: did a documented barrier get fixed after notice?
| Status | Agencies | Share |
|---|---|---|
| Corrected tested issue | 14 | 60.9% |
| Not corrected by May 2026 | 9 | 39.1% |
| Total tested | 23 | 100% |
The management reviews found partial practice adoption
GAO examined the Department of Homeland Security, Agriculture Department and Department of Housing and Urban Development in more detail. USDA and HUD had only partially followed selected leading practices for managing feedback and remediation.
GAO issued two recommendations, one to USDA and one to HUD. Neither agency agreed nor disagreed. That response is not refusal, but it leaves implementation dependent on later evidence rather than a stated commitment.
The public metric should be repair time, not policy prose
The Scoop recalculated the repair shares from GAO's counts and distinguished the tested issues from agency-wide compliance. The report does not establish that all services at the nine agencies are inaccessible or that the 14 agencies have no other barriers.
A durable measure would publish the number of verified accessibility reports, median days to remedy, retest results and unresolved high-impact barriers. That would show whether feedback systems change the service people actually encounter.
