Explainer · 2 min read
Why write down the rules of an experiment before starting?
Published on
Imagine a darts contest in which the highest-scoring areas were chosen after everyone had thrown their darts. The result would be difficult to interpret. Research faces a similar problem when its criteria for success are chosen after examining the data.
Preregistration means recording key decisions beforehand: the question, the number of participants, planned exclusions and the analyses to be used. Researchers can then compare what was announced with what was done.
Richard Wiseman and Caroline Watt compared preregistered psi research with work published without preregistration. In the groups studied, 16% of preregistered hypotheses produced a statistically significant result, compared with 46.6% of the others.
The difference is clear. However, the groups may also differ in other ways: the entire gap cannot be attributed to preregistration alone.
This practice still allows researchers to explore unexpected findings. They simply need to state that the pattern was identified afterwards, then test it using new data.
Announcing the rules helps distinguish a successful prediction from an explanation developed after seeing the result. Both can advance research, though they carry different evidential weight.
Further reading: Bulletin of 18 September 2026 (in French).
Source: Wiseman and Watt (2026), “Study preregistration: An early example and analysis”, PeerJ. Full text.