25 problems carry measured a stand-in

An institution — a regulator, a certifier, a funder, an optimizer — steers by a number that's easier to measure than the real outcome, and the two quietly diverge. Correct behavior goes uncredited; wrong behavior passes the test; sometimes people learn to game the number itself. Example: workplace noise rules rate earplugs by laboratory tests, while the sound dose actually reaching workers' ears is never measured — so millions of 'protected' workers keep losing their hearing. The problem persists precisely because the official scoreboard says everything is fine.