An evaluation of the permutation method in testing the difference between multivariate matched pairs

The impossibility of sampling a new sample or increasing the sample size of a single sample randomly sampled from a population with an unknown distribution is a frequent case faced with by researches in real applications. In certain structures of hypothesis, if all the possible permutations of the units of such a sample can be attained, the distribution free sample statistics and the decision rules of testing can be attained, the distribution free sample statistics and the decision rules of testing can be defined by the researcher within the framework of the fundamental logic of statistical hypothesis testing. This approach too, which is called as a permutation testing method, is one of the statistical testing techniques with a theoretical base established in the past with come to life a new with the rapid development in the computer technology. In this work, after giving the fundamental information and the examples, the performance of the permutation method in the structure of a hypothesis testing which aims at determine the significance of difference between the multidimensional matched pairs has been vi presented by a Monte Carlo experiment realized in the virtual environment.