This paper reviews aspects of utilizing various research scales (polyscale method) in spatial heterogeneity studies. Analysis is based on socio-economic data for different subdivision levels in Russia – municipalities, regions and economic regions (‘rayon’). Operating on lower levels of spatial organization, in accordance with fundamentals of statistics theory, is determined as the source of an increase in the ratio of observed heterogeneity. Yet, the loss of information on the rate of heterogeneity is observed when applying polyscale method. Such inaccuracies occur owing to the continuity of space and artificial nature of geographical borders. Generalization ratio which defines the difference of heterogeneity indices on spatial scales is proposed as a measure of polyscale heterogeneity. The ratio allows to distinguish the scale levels with the highest variety of spatial units, along with the degree of actual geographical contribution to the heterogeneity indicators. Calculations example on data for microregions, as proposed by E. Leizerovich, shows considerable differences between actual and assessed heterogeneity. Application of the method in research studies is proposed, e.g. typology assessment or multi-level forecasting.

No abstract