In the course of preparing materials to explain boxplots to students, I am seeing something odd with the whiskers in a boxplot. The code
## example box
X = randn(100)
M = median(X)
Q1 = quantile(X,0.25)
Q3 = quantile(X,0.75)
IQR = iqr(X)
boxplot(X,label=:none,title="example boxplot", outliers=false)
scatter!([1.], [M],label="median")
scatter!([1.], [Q1],label="Q1")
scatter!([1.], [Q3],label="Q3")
scatter!([1.], [Q1-1.5*IQR],label="Q1-1.5IQR")
scatter!([1.], [Q3+1.5*IQR],label="Q3+1.5IQR"
generates the plot
What I do not understand is why the points labeled Q1-1.5IQR and Q3+1.5IQR do not lie exactly at the same height as the whiskers. The boxplot code in StatsPlots.jl uses 1.5 as the default whisker_range.