Analysing How People Orient to and Spread Rumours in Social Media by Looking at Conversational Threads (1511.07487v3)

Abstract: As breaking news unfolds people increasingly rely on social media to stay abreast of the latest updates. The use of social media in such situations comes with the caveat that new information being released piecemeal may encourage rumours, many of which remain unverified long after their point of release. Little is known, however, about the dynamics of the life cycle of a social media rumour. In this paper we present a methodology that has enabled us to collect, identify and annotate a dataset of 330 rumour threads (4,842 tweets) associated with 9 newsworthy events. We analyse this dataset to understand how users spread, support, or deny rumours that are later proven true or false, by distinguishing two levels of status in a rumour life cycle i.e., before and after its veracity status is resolved. The identification of rumours associated with each event, as well as the tweet that resolved each rumour as true or false, was performed by a team of journalists who tracked the events in real time. Our study shows that rumours that are ultimately proven true tend to be resolved faster than those that turn out to be false. Whilst one can readily see users denying rumours once they have been debunked, users appear to be less capable of distinguishing true from false rumours when their veracity remains in question. In fact, we show that the prevalent tendency for users is to support every unverified rumour. We also analyse the role of different types of users, finding that highly reputable users such as news organisations endeavour to post well-grounded statements, which appear to be certain and accompanied by evidence. Nevertheless, these often prove to be unverified pieces of information that give rise to false rumours. Our study reinforces the need for developing robust machine learning techniques that can provide assistance for assessing the veracity of rumours.

• The paper introduces a detailed annotation scheme for 330 rumour threads, revealing that true rumours resolve in 2 hours versus over 14 hours for false ones.
• It finds that unverified rumours gain an initial surge of retweets while supportive responses dominate until debunking occurs.
• The study highlights the influence of high-reputation users and suggests potential for real-time rumour verification using machine learning.

Analysing How People Orient to and Spread Rumours in Social Media by Looking at Conversational Threads

This paper presents an incisive exploration of the dynamics surrounding rumours on social media, specifically focusing on how these rumours are initiated, supported, or debunked within conversational threads. The researchers developed a robust methodology to curate and examine a dataset encompassing 330 rumour threads (comprising 4,842 tweets) across nine significant events.

Key Findings

1. Resolution Time of Rumours: True rumours are resolved more quickly than false ones, with median resolution times of approximately 2 hours for true rumours compared to over 14 hours for false rumours. This aligns with the hypothesis that disproving a statement often takes longer than corroborating it.
2. Rumour Propagation Patterns: The propagation analysis reveals that unverified rumours generate an initial surge of retweets, indicating heightened interest at onset. This interest wanes for both confirmed and debunked rumours, with the retweet intensity peaking within the first few minutes post-release.
3. Support vs. Denial: Users predominantly support unverified rumours, regardless of their eventual truth value. Notably, denial tweets surpass supportive ones only when false rumours are debunked. This suggests that the community lacks skepticism in the absence of counter-evidence.
4. Certainty and Evidentiality: Certainty levels exhibited minimal variance across rumour statuses. In contrast, evidentiality decreased post-resolution, particularly for false rumours. There was a notable drop in evidence-based tweets after rumour debunking, shifting the focus from disproving the rumour to discussion.
5. Role of User Reputation: Users with high follow ratios, typically news organisations, often provide support and evidence for rumours, whether ultimately proven true or false. This suggests they strive for credibility, quoting sources to maintain professional standards, despite the pressure to publish promptly.

Methodological Contributions

The researchers designed a detailed annotation scheme to capture conversational attributes in social media threads, focusing on support, certainty, and evidentiality. This framework allowed for a nuanced exploration of user interactions with rumours and the gathering of significant insights into user behaviour.

Implications

The findings underscore the challenge of verifying information in real-time during unfolding news events. This paper reinforces the need for advanced machine learning models capable of assessing rumour veracity dynamically. Such developments could assist platforms and journalists in mitigating the spread of misinformation.

Future Directions

The potential for leveraging these findings in machine learning applications is significant. Future research could focus on real-time rumour detection and verification models, potentially incorporating the identified conversational patterns as features for better automated veracity assessments.

In conclusion, this work provides a comprehensive analysis of rumour dynamics on Twitter, offering valuable perspectives for enhancing the verification processes and improving the reliability of information dissemination on social media platforms.