Large sample sizes With large enough sample sizes, we’ll always reject, so why perform the test? Previous Ignoring the alternative Next Deductive invalidity