Methodology for the Study of Theoretical Narratives Using Computer Text Analysis

This article is devoted to the problem of studying theoretical narratives in economics by analyzing textual information using natural language processing methods. Generally accepted approaches for conducting such an analysis exist in the social sciences, in particular in political science. However, within the framework of economic research related to the analysis of textual information, the methodology for solving this problem is only at the beginning of its development path. The purpose of this review article is to analyze related works on this topic, as well as to propose a methodology that would solve the problem of identifying ideological narratives not only using quantitative methods, but also using qualitative analysis and "semantic aggregation" of textual information. Based on the analysis of Russian economics, a set of practical steps is proposed to identify ideological plots, including the use of deep learning models and text semantics aggregation. The proposed approach can contribute to a more accurate identification and interpretation of ideological constructs of economic theory, which opens up new prospects for critical analysis and the development of economics.