ABSTRACT
With the fall in the price of personal computers along with the concurrent increase in their processing speed, randomization tests have started to gain in their popularity among statisticians. The value of randomization tests lies in their wide applicability and their application to problems for which there is no parametric solution. Specific types of randomization tests are given names such as the bootstrap or jackknife, but some are simply called permutation tests. Surprisingly scientists in other fields have failed to grasp and make use of them. But just what are permutation or randomization tests? These tests are the generic names applied to computer-intensive methods that generate either the sampling distribution, the standard error, or p-value of a test statistic.
