This is the introductory article of a series of eight where the authors discuss non-parametric methods. They compare the advantages and disadvantages of non-parametric methods with parametric ones. Non-parametric methods can be used in particular where the investigated variables do not have a normal distribution, where the small size of the investigation and the type of classification do not permit to test, when information on the investigated variables are presented by means of rank or normal scales etc. Their disadvantage is a smaller ability to refute the null hypothesis, when it is not correct, as compared with parametric tests.

Katedra sociálního lékarství fakulty Masarykovy univerzity Brno