A way of summarising data distribution. This technique divides up the range of possible values in a data set into intervals, such that each interval contains approximately the same percentage of the values. A set of statistics are collected for each interval.