RESEARCH ON THE NORMAL DISTRIBUTION OF VARIABLES AS A CONDITION FOR THE CORRECT APPLICATION OF THE STATISTICAL MODELS FOR PROCESSING DATA

The paper aimed to test the normality of the distribution of the variables using Pearson's central moment coefficients. The data were referring to the milk and meat production characters collected from 5,817 individuals, offspring from 106 Frisian bulls under breeding value estimation. The variables were used both in terms of metric characters and logarithmic characters. In case of milk production and reproduction traits, the use of logarithmic characters was useless, and the values of the Beta coefficients showed a distribution relatively closer to the normal one. In case of meat production characters, both the use of metric and logarithmic characters reflected that the distribution of variables was not a normal one, on the contrary it was sharply vaulted. As a conclusion, it is important to check the normality of the variables distribution in order to avoid the errors of calculation by eliminating the extreme values in case of the skewed distribution.