ABSTRACT
Tests are classified either according to their objective (goal) or their mathematical properties.
A test is considered to be a parametric test if its objective is to test certain hypotheses related to one or more parameters of a random variable of a specified distribution law. In most cases, these tests are based on the fact that the data series follows a normal distribution and assume that the random variable of reference X follows a normal distribution. The question is to find out whether the results remain valid even if the distribution of X is not normal: if the results are valid we say that the test in question is robust. The robustness of a test compared to a model is its ability to remain relatively insensitive to certain modifications of the model. A test is known as nonparametric if it does not call upon parameters or precise hypotheses concerning the subjacent distribution.
