That is, all experiments with P -values that did not reach statistical significance are discarded and a particular P -interval is constructed only if the obtained P -value is close to 0.05.
← all excerpts
Bayesian prediction intervals for assessing P-value variability in prospective replication studies.
1
—
—