Reconstruction of Media Social Representations Using Indicators of Text Statistics (Based on Media Discourse on the Pandemic)

Abstract

Objective. The aim is to present an algorithm to reconstruct media social representations based on indicators of text statistics and to conduct a comparative analysis of the construction of semantically similar media joint ventures, such as “pandemic”, “coronavirus”, “COVID-19” in Russian media. Background. Social representations perform the most important functions in the process of social functioning of an individual and a group, serve as a tool for cognition, adaptation and regulation of behavior and are formed taking into account the influence of media social representations. Methods for studying social representations for various social groups are presented in psychological studies, however, methods for studying media social representations are discussed in single scientific work. The presented scientific project is based on the theory of social representations by S. Moskovici and generalizations by B. Hoyer regarding the construction of media social representations (naming, emotional attachment, thematic attachment, metaphorical attachment and attachment through basic antinomies). Study design. The phenomenon of the coronavirus pandemic, presented in media discourse, was used as the signified in the study. The signifier is a trio of semantically similar concepts (“pandemic”, “COVID”, “coronavirus”). Measurements. To reconstruct media social representations, statistically stable collocations were identified to indicate the measure of association, logically close to the associative experiment. Hence, it was possible to identify thematic networks, axiological and evaluative components, components-characteristics of activity. The research material is represented by texts about the COVID-19 pandemic (January 2020-March 2022: “Rossiyskaya Gazeta”: 19471 texts, 7,97 million words; “Kommersant”: 1482 texts, 1,07 million words, “Novaya Gazeta”: 705 texts, 0,9 million words) and processed using BootCat, TreeTagger, AntConc (lemmatization, frequency analysis). Results. The associative fields of the joint ventures are different and contain anchoring and objectification resources when using concepts, only few elements of the associative field are similar. Thus, depending on the concept used to signify the intent of the text, stories about the fear of infection, the treatment of the disease, or resistance to the harsh elements are created. Conclusions. The similar algorithm based on media text statistics can be used to reconstruct any media social representation.