Diana Bílková, Faculty of Informatics and Statistics, University of Economic in Prague, Czech Republic

Pages: 35 – 56

Abstract

This paper deals with the construction of wage distribution models using fourparameter and three-parameter lognormal curves. The main objective of the research
is to compare the accuracy of the models obtained using both types of lognormal
distribution as wage distribution models. The minimum wage in a given year represents the beginning of four-parameter and three-parameter lognormal curves. The estimates of the remaining three or two parameters are constructed using the quantile method. The chi-square testing criterion is used for the evaluation of the accuracy of the models obtained. In almost all cases of wage models, the fourparameter lognormal curves yielded more accurate results than the three-parameter lognormal curves. The results in terms of accuracy of both types of lognormal curves are equally accurate only in a few cases. However, the differences in the accuracy of the four-parameter and three-parameter wage models are not critical.