3 ms·
Sounds like the old p=0.05 problem!
by quickthrower2 7y ago
Sounds like the old p=0.05 problem!
- roel_v 7y agoOK so I undown'ed you to ask - how?
- quickthrower2 7y agoImagine you have 20 experiments, each with a 5% chance that the null hypothesis shows a positive result, then there is a good chance that one of them will be positive even if null is true for all of them. That experiment is bunk but you say p=0.05 so all is good!