Why should a Shapiro–Wilk p-value for a very large sample be interpreted cautiously?
StatisticsLevel: IntermediateDistribution checking
Why should a Shapiro–Wilk p-value for a very large sample be interpreted cautiously?
Hints
- Check the implementation's sample-size note.
- Statistical detectability is not practical importance.
Answer
The documented approximation can lose accuracy beyond its recommended range; large samples also detect tiny departures. Examine plots and consequences for the downstream procedure.
Source and licence: SciPy — distribution diagnostics and robust tests · BSD-3-Clause · answer: Quantoros editorial