The influence of classification method selection on the identification of spatial dependence – application of join-count test

A lot of regional studies cover classification of territorial units due to considering research problem. This approach examines territorial diversification of a phenomena, as well as spatial interactions. The occurrence of spatial dependence can reveal the processes of creating or extanding spatial clusters. One of the significant determinants of research results is a way of classification. Different approach leads to diversified divisions of territorial units. The objective of this paper is to examine the influence of classification method selection on the spatial autocorrelation analysis results using join-count test. This test, in contradiction to the other methods proposed in the field of spatial statistics, analyzes spatial autocorrelation based on qualitative data. Therefore, it can be applied in the examination of spatial dependence between territorial units from distinguished classes of territorial units.