Echo Chambers and Science: Analysing the Impact of Digital Homogenisation on Public Discourse over Science
Echo Chambers and Science: Analysing the Impact of Digital Homogenisation on Public Discourse over Science
Author(s): Svetlomir ZdravkovSubject(s): Social Sciences, Sociology, Social Theory, Social Informatics, Sociology of Culture, Sociology of the arts, business, education
Published by: Институт по философия и социология при БАН
Keywords: science polarisation; public understanding of science; echo chambers; topic modeling; social homophily
Summary/Abstract: This study examines the evolution of homogenisation in scientific discourse on digital platforms in Bulgaria from 2011 to 2022 by analysing a comprehensive textual corpus of digital media. Using BERTopic, an embedded topic modeling technique, the analysis identified key topics that drive both polarisation and consensus. By applying a polarization index (PI) to measure the variance in emotional responses to science-related topics, the analysis reveals a shift from diverse, high-polarisation discussions in the early years of the platform to more homogeneous, low-polarisation interactions as the platform matured. Initially, topics such as space exploration and education systems generated significant public debate, reflecting a diverse engagement with science. Over time, however, the platform became more insular, with discussions increasingly confined to an echo chamber that reinforced existing beliefs, particularly in soft sciences such as education. The findings suggest that, while digital platforms may initially foster diverse scientific discourse, their design and resulting homogenising tendencies ultimately lead to greater polarisation and fragmentation in the public sphere. This not only distorts the public’s understanding of science, but also undermines society’s ability to address complex scientific challenges effectively, as is required to address multiple fragmented publics rather than a single one.
Journal: Социологически проблеми
- Issue Year: 56/2024
- Issue No: Special
- Page Range: 59-78
- Page Count: 20
- Language: English
- Content File-PDF