The spatial weights matrices and their influence on the quality of spatial models of employment

In the article the problem of employment modelling is analysed. The analyses have been carried out taking into account the spatial factor. To take into account of this factor means to use the spatial weights matrices, that define the spatial structure of the neighbourhood and that are an essential part of the research. The studies included regions of selected countries of Central Europe. The role of the explained variables played the characteristics describing the number of employees in total and the number of employed men and women, while the explanatory variables were selected demographic characteristics and GDP per capita. The studies included two basic spatial models: spatial lag model and spatial error model - in each case, a better tool of analysis turned out to be the spatial error model. In the analyses different spatial weights matrices were considered and the best of the variants were indicated.