Non-parametric statistics
Non-parametric statistics are methods that test hypotheses without assuming the data follow a specific distribution, typically by analysing the ranks of the observations.
See it move
Loading infographic...
Non-parametric statistics test a hypothesis without assuming a specific distribution, typically by ranking the data first. Five branch managers' bonuses of €400, €420, €430, €450 and €5,000 become ranks 1 through 5 once sorted, so the one extreme €5,000 payout influences the result only by one rank position, not by thousands of euros. This makes rank-based tests the fallback whenever a sample is small or clearly skewed.
Where it fits
SubjectData Analysis & StatisticsAdvancedTopicHypothesis TestingAdvancedTopicCommon Significance TestsAdvanced