The absurd but true notion that a coin flip provides approximately the same positive predictive value as an exact replication experiment when the initial effect is barely significant highlights an important, but little known, limitation of exact replications.
1
—
—