randomization test
an inferential approach that combines observed data across all participants and experimental conditions and then randomly sorts the data into new samples (groups). A test of statistical significance is performed, and the value obtained is compared with the value that was obtained when the data were in their original form. This process is repeated many times. For example, suppose there were 1,000 recalculated mean differences between the experimental (E) and control (C) groups, derived from 1,000 permutations. Comparison of the 1,000 recalculated means to the original means provides an exact probability of getting the original means. If 25% of the mean differences between the E and C group permutations differ by as much as or more than the difference between the original E and C means, then the probability value for the original mean difference is p = .25. A randomization test is a nonparametric approach; that is, it does not make assumptions about
the distribution of the data. Also called permutation test.