2 ms·
Just a reminder that if 20 interested things are studied, on average you should expect 1 of them to be a false positive with p<0.05. It's a bit like throwing da
by nwallin 5y ago
Just a reminder that if 20 interested things are studied, on average you should expect 1 of them to be a false positive with p<0.05. It's a bit like throwing darts: it helps if there exists a relationship, but it also helps to be lucky.
relevant xkcd: https://xkcd.com/882/ https://xkcd.com/882/
- AussieWog93 5y agoIn theory, it's 1 in 20. In practice, it's much, much higher, closer to 2 out of 3: https://en.m.wikipedia.org/wiki/Replication_crisis https://en.m.wikipedia.org/wiki/Replication_crisis