Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

What Makes A Video Radicalizing? Identifying Sources of Influence in QAnon Videos

Lin Ai, Yu-Wen Chen, Yuwen Yu, Seoyoung Kweon, Julia Hirschberg, Sarah Ita Levitan

TL;DR

Radicalization increasingly occurs on video platforms; this paper investigates how viewer traits and multimodal video cues influence perceptions of QAnon videos. It builds a large corpus (5,924 videos) and a comprehensive questionnaire to derive three perception metrics (Enjoyment, Content, Actions) and to identify influential viewer traits and video features. The authors analyze correlations between viewer traits and responses and examine textual, acoustic, and visual cues at segment and video levels, uncovering patterns such as violence-related language reducing positive perception and weapon presence diminishing perceived quality. These findings provide a data-backed basis for understanding radical-content appeal and inform future multimodal detection and moderation efforts.

Abstract

In recent years, radicalization is being increasingly attempted on video-sharing platforms. Previous studies have been proposed to identify online radicalization using generic social context analysis, without taking into account comprehensive viewer traits and how those can affect viewers' perception of radicalizing content. To address the challenge, we examine QAnon, a conspiracy-based radicalizing group, and have designed a comprehensive questionnaire aiming to understand viewers' perceptions of QAnon videos. We outline the traits of viewers that QAnon videos are the most appealing to, and identify influential factors that impact viewers' perception of the videos.

What Makes A Video Radicalizing? Identifying Sources of Influence in QAnon Videos

TL;DR

Radicalization increasingly occurs on video platforms; this paper investigates how viewer traits and multimodal video cues influence perceptions of QAnon videos. It builds a large corpus (5,924 videos) and a comprehensive questionnaire to derive three perception metrics (Enjoyment, Content, Actions) and to identify influential viewer traits and video features. The authors analyze correlations between viewer traits and responses and examine textual, acoustic, and visual cues at segment and video levels, uncovering patterns such as violence-related language reducing positive perception and weapon presence diminishing perceived quality. These findings provide a data-backed basis for understanding radical-content appeal and inform future multimodal detection and moderation efforts.

Abstract

In recent years, radicalization is being increasingly attempted on video-sharing platforms. Previous studies have been proposed to identify online radicalization using generic social context analysis, without taking into account comprehensive viewer traits and how those can affect viewers' perception of radicalizing content. To address the challenge, we examine QAnon, a conspiracy-based radicalizing group, and have designed a comprehensive questionnaire aiming to understand viewers' perceptions of QAnon videos. We outline the traits of viewers that QAnon videos are the most appealing to, and identify influential factors that impact viewers' perception of the videos.
Paper Structure (23 sections, 4 figures, 29 tables)

This paper contains 23 sections, 4 figures, 29 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Corpus and Annotation Collection
  4. Rater Demographics and Background
  5. Evaluation Metrics
  6. Analyzing Viewer Ratings and Traits
  7. Analysis of Video Characteristics
  8. Multimodal Feature Analysis
  9. Textual Features
  10. Acoustic Features
  11. Visual Features
  12. Conclusions and Future Work
  13. Significant Multimodal Features
  14. Segment Level Significant Features
  15. Textual Features
  16. ...and 8 more sections

Figures (4)

  • Figure 1: Rater demographics. A total of 46 raters completed the questionnaire.
  • Figure 2: Rater self-reported personalities. A total of 46 raters completed the questionnaire.
  • Figure 3: Rater's opinion on radical groups. A total of 46 raters completed the questionnaire.
  • Figure 4: Rater's opinion on media sources. A total of 46 raters completed the questionnaire.