Can median absolute deviation be zero when some observations differ?
StatisticsLevel: IntermediateRobust summaries
Can median absolute deviation be zero when some observations differ?
Hints
- Compute absolute deviations.
- Then take their median.
Answer
Yes: [0,0,0,1,100] has median zero and MAD zero. A majority of identical values masks minority dispersion.
Source and licence: SciPy — statistical estimators and hypothesis tests · BSD-3-Clause · answer: Quantoros editorial