How do trimming and winsorization treat extreme observations differently?
StatisticsLevel: IntermediateRobust summaries
How do trimming and winsorization treat extreme observations differently?
Hints
- Deletion versus replacement.
- Check the retained sample size.
Answer
Trimming removes them; winsorization replaces them with boundary values. Sample size and the resulting estimand therefore differ.
Source and licence: SciPy — statistical estimators and hypothesis tests · BSD-3-Clause · answer: Quantoros editorial