1. Use median
median(x)
The median is a summary statistic which isn't 'thrown' by outliers.
2. Use a trimmed mean
mean(x, trim = .05)
A trimmed mean is the mean of all values except those in the highest and lowest n
%, where n
is a value you choose. The code above would return the mean of every value in x
except the highest 5% and the lowest 5%.
Whatever you do, make sure you report transparently.
与恶龙缠斗过久,自身亦成为恶龙;凝视深渊过久,深渊将回以凝视…