A measure of the difference between two sets of data, often data collected from a treatment group and from a control group.