Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Echo-chambers and Idea Labs: Communication Styles on Twitter

Aleksandra Sorokovikova, Michael Becker, Ivan P. Yamshchikov

TL;DR

The paper addresses the assumption that online vaccination discourse on Twitter consists predominantly of echo chambers by exploring the diversity of communication styles across communities. It combines graph-based community detection (OpenOrd layout and Louvain modularity) on reply networks with text analytics (RoBERTa TweetEval polarity, TextBlob subjectivity, and a logical-fallacy detector) to profile each community’s discourse. The study identifies six dense vaccine-related communities with distinct textual and interaction patterns, and shows that textual features can partially predict community membership (accuracy $0.35$, rising to $0.44$ after excluding a weakly connected cluster), with a modularity of $0.902$ for the detected communities. This work highlights the need for a nuanced taxonomy of online discourse beyond echo chambers and provides a computational framework for assessing and labeling online communities in health-related discussions.

Abstract

This paper investigates the communication styles and structures of Twitter (X) communities within the vaccination context. While mainstream research primarily focuses on the echo-chamber phenomenon, wherein certain ideas are reinforced and participants are isolated from opposing opinions, this study reveals the presence of diverse communication styles across various communities. In addition to the communities exhibiting echo-chamber behavior, this research uncovers communities with distinct communication patterns. By shedding light on the nuanced nature of communication within social networks, this study emphasizes the significance of understanding the diversity of perspectives within online communities.

Echo-chambers and Idea Labs: Communication Styles on Twitter

TL;DR

The paper addresses the assumption that online vaccination discourse on Twitter consists predominantly of echo chambers by exploring the diversity of communication styles across communities. It combines graph-based community detection (OpenOrd layout and Louvain modularity) on reply networks with text analytics (RoBERTa TweetEval polarity, TextBlob subjectivity, and a logical-fallacy detector) to profile each community’s discourse. The study identifies six dense vaccine-related communities with distinct textual and interaction patterns, and shows that textual features can partially predict community membership (accuracy , rising to after excluding a weakly connected cluster), with a modularity of for the detected communities. This work highlights the need for a nuanced taxonomy of online discourse beyond echo chambers and provides a computational framework for assessing and labeling online communities in health-related discussions.

Abstract

This paper investigates the communication styles and structures of Twitter (X) communities within the vaccination context. While mainstream research primarily focuses on the echo-chamber phenomenon, wherein certain ideas are reinforced and participants are isolated from opposing opinions, this study reveals the presence of diverse communication styles across various communities. In addition to the communities exhibiting echo-chamber behavior, this research uncovers communities with distinct communication patterns. By shedding light on the nuanced nature of communication within social networks, this study emphasizes the significance of understanding the diversity of perspectives within online communities.
Paper Structure (12 sections, 2 figures, 2 tables)

This paper contains 12 sections, 2 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Data
  3. Experiments
  4. Community clustering
  5. Text Classification Metrics
  6. Polarity Scores
  7. Subjectivity
  8. Logical fallacy
  9. Predicting Community Membership with Communication Style
  10. Discussion
  11. Conclusion
  12. Bibliographical References

Figures (2)

  • Figure 1: Conversation graph after applying OpenOrd algorithm
  • Figure 2: Average scores for mean subjectivity and mean negativity in all 6 user communities. One can clearly see that one of the community is characterised as highly subjective while another is highly negative while scoring lower on subjectivity.