TREATMENT OF THE DATA FILES WHICH DO NOT MEET BASIC CONDITIONS OF SELECTED STATISTICAL METHOD

The authors of the paper try to point out that there is one important problem in the basic statistical methods teaching at faculties of economics. The basic courses of statistics usually contain just knowledge about the statistical methods which use is connected with some requirements, especially with the data normality or homoscedasticity etc. But many students are surprised when they work with real data files and they find out that the data have "unsuitable" characteristics in relation to the requirements of selected methods. The aim of this paper is to present possible ways in such cases when some conditions of selected methods cannot be satisfied. We focused on the methods which are usually taught in faculties of economics in the frame of two-semester course of statistics. We mean the parametric hypotheses testing and analysis of variance above all and other methods. At the same time, we also discuss the assumptions of the methods use which can replace the basic methods. The truth is that these substitute methods are also connected with some assumptions.