The original finding was small, striking and endlessly repeatable in headlines. The replication was large, pre-registered and null. Only one of them shaped a decade of public advice.

Why null results travel badly

A study that finds nothing is harder to summarise, harder to publish and far harder to headline. The incentive structure quietly favours the surprising result over the sturdy one.

What changed

Pre-registration and registered reports have made the problem visible, which is not the same as fixing it. The correction still takes years to reach practice.