Explaining permutation tests
When observing a phenomenon belonging to a set of possible results, we ask ourselves what the law of probability is that we can assign to this set. Statistical tests provide a rule that allows us to decide whether to reject a hypothesis based on the sample observations.
Parametric approaches are very uncertain about the experiment plan and the population model. When these assumptions are not respected, particularly when the data law does not conform to the needs of the test, the parametric results are less reliable. When the hypothesis is not based on knowledge of the data distribution and assumptions have not been verified, nonparametric tests are used. Nonparametric tests offer a very important alternative since they need fewer hypotheses.
Permutation tests are a special case of randomization tests that use series of random numbers formulated from statistical inferences. The computing power of modern computers has made their widespread application...
