In science, what level of probability is required before a new discovery is accepted?

Question

When discovering a new particle or exoplanet or anything, a spike in the data must be higher than a random fluctuation in the background noise. How many standard deviations above average must it be for scientists to accept it as a discovery? What percent of uncertainty is acceptable?

Dixon · Answer

It isn't as binary as that. A discovery/idea will have a certain plausibility based how reasonable it seems in the context of its field and how much experimental evidence there is for it, including the quality of that evidence. Some things that match expectations but only have, say, 95% statistical confidence from one experiment are often accepted as true. Whereas extraordinary ideas that run counter to the mainstream can require a lot of very high quality evidence to get accepted.

Note that while science itself is about reason and evidence, acceptance is a human trait and scientists are still humans with their own preferences.

Note that while science itself is about reason and evidence,  acceptance is a human trait and scientists are still humans with their own preferences.

? · Answer

Everyone can have their own opinion .

οικος · Answer

Generally, the cutoff point is 95% probability (5% chance of a random event). 99% is better.

In science, what level of probability is required before a new discovery is accepted?

3 Answers