Overview for Friedman Test

Use Friedman Test to determine whether the median treatment effect differs in a randomized block design.

For example, a pharmaceutical researcher wants to evaluate the median effects of three drug therapies on enzyme activity in animals from a set of litters. The researcher uses the litters as blocks because she wants to account for the genetic differences between litters.

With this test, you must have the following:
  • The response variable must be continuous or ordinal.
  • The treatment must be categorical.
  • You must have one categorical block variable. Blocks are a group of experimental runs that are conducted under consistent conditions. When you account for the effects of the block variable, you can assess the treatment effect more precisely.
The Friedman test is a nonparametric alternative to a DOE model with blocks, and to an ANOVA model with two factors.

Where to find this analysis

To perform the Friedman test, choose Stat > Nonparametrics > Friedman.

When to use an alternate analysis

If you have only one categorical factor and nonnormal data, use Kruskal-Wallis Test or Mood’s Median Test.