4 ms·

Shouldn't we want replication even if the sample size seems large, to rule out experimental errors or other flaws in the work?