Changes in the Methodology of Socio-Economic Research under the Impact of Advanced Technologies

Selected aspects of methodological developments in socio-economic research are addressed from a conceptual perspective, with outlining the areas of transformations in methodology, and illustrating the feasibilities of its employment in present-day analytical and forecasting developments. The source base of this study is open analytical documents, reports, methodological and working reviews, and political documents of the Organization for Economic Cooperation and Development. Relevant scientific developments of various thematic focus were selected by criteria of novelty of technical toolkit, employment of alternative or high-frequency data sources, orientation on applied results for policy setting. An analysis of these materials enabled the author to identify a series of important trends in the methodological developments of socio-economic research, and to show the ways of combining traditional and novel approaches. Technical contained in reports on research results descriptions were used; specific methods were examined, applicable in descriptive diagnostics, identification of causalities, prompt assessment of the current performance, forecasting, and scenario-based modelling. Special emphasis was placed on the integration of algorithmic models with classical statistical and econometric tools, which a central feature of the current methodological transformation. The analysis demonstrated that the methodology of current socio-economic research was undergoing deep changes caused by digitalization, increasing scopes of available data, and the increasing complexity of research objects. There is a transition from the prevalence of classical statistical techniques to ever extending applications of flexible combined techniques of machine learning, big data analysis, scenario-based forecasting. This enables for fuller accounting for non-linear relationships, for more meticulous exploring a wide range of factors and premises affecting the development of socio-economic systems. The employment of algorithmic models and prompt assessment methods reduces the dependence on insufficiently flexible statistical information, and increases the accuracy of analytical conclusions and forecasts.