New relevance and significance measures to replace p-values.

Clearly, using the criterion of a secured relevant effect (case Rlv) as a filter would reduce the frequency of phony results drastically, since the barely significant results would rarely pass it.

also in 875 other papers