Low values of NPI-B-RP are worrying from a practical perspective, in particular in case H 0 is rejected with the p -value only just below the level of significance, because many experiments are explicitly designed with the aim to find evidence supporting H 1 .
← all excerpts
Statistical reproducibility for pairwise <i>t</i>-tests in pharmaceutical research.
2
—
—
The sentences
The key observations are: NPI-B-RP is low when the p -value is close to the level of significance α .