Friedman test

The Friedman test addresses the issue of deciding whether k matched samples were drawn from the same population. It is therefore an identity test.

The observations must be measured on an numerical or ordinal scale (i.e. not categorical).

The Friedman test is non parametric: it does not make any assumption on the underlying distributions. As many non parametric tests, it will first convert the raw values into ranks, and then build an appropriate statistic from these ranks.

The Friedman test should not be confused with the Kruskal-Wallis test, that deals with the issue of whether k independent samples were drawn from the same population.

____________________________________________

Related readings :