Question

I noticed that median of boxplot (constructed with restricted ylim-parameter) may differ from the median obtained by median()-function or boxplot without adjusted y-axis.

Does it mean that boxplot use for calculations only values lying within defined interval of y-axis??
And if so, how could I get correct boxplot (based on all values), but plot it on particular interval of y.axis??

Thank you very much.

Answer 1

Using ylim restricts the scale for y. In ggplot, data that falls outside the scale’s limits is discarded and is not included in any statistical computations (such as boxplots). To just zoom the axes without changing the data that is included, use coord_cartesian(ylim=...).