A statistic is mean(A)−mean(B). Which alternative tests whether A is larger?
StatisticsLevel: IntermediatePermutation experiment design
A statistic is mean(A)−mean(B). Which alternative tests whether A is larger?
Hints
- Track the sign of the contrast.
- The right tail supports a positive effect.
Answer
Use greater: count null statistics at least as large as observed. Reversing the statistic without reversing the alternative tests the opposite claim.
Source and licence: SciPy — resampling and Monte Carlo inference · BSD-3-Clause · answer: Quantoros editorial