Science Puzzle
The Rat Tail Bounty
In 1902 the French colonial government in Hanoi, Vietnam, feared an outbreak of plague, a disease spread by fleas on rats. Rats were thriving in the city's new sewers, so the authorities offered a reward: one cent for every rat tail handed in at a police station. Tails poured in, with more than 20,000 on a single day in June.
Yet the city's rat problem got no better, and people began to spot rats scurrying around with no tails. What had gone wrong?
The Answer
The reward was for tails, not for dead rats. So people caught rats, cut off their tails and let them go to breed more rats.
A tail was supposed to be proof that a rat had been killed. But a rat can live perfectly well without its tail, and a live rat is worth far more to a rat catcher than a dead one: it can have babies, each with a tail worth a cent. Some people went further and set up rat farms on the edge of the city to breed rats for their tails. The authorities had created a business that depended on keeping the rats alive. The scheme was dropped.
This trap is known as Goodhart's law, often put as "when a measure becomes a target, it ceases to be a good measure." The number of tails was meant to measure progress towards fewer rats. Once people were paid for the number itself, they found the cheapest way to raise it, which had nothing to do with the real goal.
Science runs into this too. A study measured only by how many papers it produces, or a test judged only by its score, invites the same shortcuts. The safest measures are the ones closest to the thing you actually care about, and hardest to fake.
The principle: Goodhart's law. When a measurement becomes a target, people find ways to change the number without achieving the goal it was meant to track, so measure the outcome you actually want.