The ±2σ outlier rule assumes a Gaussian distribution
I always defined outliers by values laying 2 standard deviations from the mean, but it turns out it assumes a normal distribution. A heavy-tailed distribution paints a whole different picture.
Rests on
- Outliers mask themselves under ±2σ detection because they inflate the σ how this rule fails even when data is following a Gaussian distribution but contains outliers