Distribution-free statistics, statistics designed to test differences in data that either does not meet tests of normality, or can not be tested for normality because it is not ordinal or continuous (e.g. categorical data).