Дети и образовательные ресурсы: кластеризация российских регионов

Abstract

A regional education system is a complex dynamic system that is closely linked with the external social environment. The paper presents the results of a study focused on the conditions required for further development of regional secondary education systems. \nAt the first stage of the research, the author designated clusters for Russian regions according to eight criteria. The Childhood Social Potential in Russian Regions database was used as the data source. The database contains more than 150 management and target factors that describe the results of primary, junior secondary and senior secondary education programs, as well as socio-economic, demographic, infrastructural, natural-environmental and other indicators for 85 subjects of the Russian Federation. The database also contains summary data from information and analytical reports made by regional information processing centres, regional centres for education quality assessment, and analytical collections produced by the Higher School of Economics National Research University. \nAll database values were analysed for outliers and extreme values by means of Deductor Studio, a special software product, and an analysis of data quality was also conducted. After the data was arranged in proper form, a number of experiments were performed to obtain a more accurate result of cluster analysis. \nThe following indicators were used as cluster-forming factors: index of pre-school education quality; index of secondary education quality; index of additional education quality; index of vocational training quality; population density, people per square km; ratio of gross regional product per capita to the cost of a fixed set of consumer goods and services, units; share of education expenditures in the consolidated budgets of the subjects of the Russian Federation and territorial state extra-budgetary funds, %; ratio of the state debt volume of the subject to the revenues of the consolidated budget of the subject and territorial state non-budgetary funds (excluding subsidies for budget equalisation), %. As a result of applying the EM-clustering algorithm, the author identified five clusters. When describing cluster profiles, the author determined the power, significance, and composition of the clusters and evaluated and analysed the average values of cluster-forming factors and the average values for all 85 regions of the country.