Exploratory study on the dynamics of online responses to debunking messages about covid-19 in Singapore on Whatsapp

Abstract

The COVID-19 pandemic had spurred a concurrent outbreak of false information online. Debunking false information pertaining to a health crisis is critical as it can affect the public’s decisions to adhere to behaviours like mask-wearing. This exploratory study examined the effects of debunking messages on a COVID-19-related public chat on WhatsApp in Singapore.

As debunking involves persuading others that another piece of information is false, the persuasiveness of these debunking messages can differ based on their degree of source credibility, such as providing evidence from authoritative sources. Hence, this study examined the relationship between source credibility of different debunking message types (official communications, organic debunking, doubting the veracity of the claims and debunking with links which also includes six tiers of differences in the URLs) and its effects on the length of the conversation, sentiments towards various aspects of a crisis, and the information distortions in a message thread. This was done by applying a mix of deep learning techniques, knowledge graphs, and content analysis to perform aspect-term sentiment analysis of the messages and measure information distortion.

The results indicated that debunking messages with high source credibility have an impact on closing a discussion thread and getting the participants to move to a different topic. It also showed that there are shifts in sentiments towards some aspects of the crisis, highlighting the potential value of Aspect-based Sentiment Analysis (ABSA) in monitoring the effectiveness of debunking messages, which was a relatively unexplored area in the literature. Finally, it was also found that debunking messages with low source credibility were likely to increase information distortion in conversation threads.

The study supports the importance of source credibility in debunking as well as an aspect-based approach analysing the effect of debunking messages, which could have practical value for public agencies during a health crisis and contribute to the rumour and debunking literature. Additionally, this study contributes a novel approach to measuring information distortion using knowledge graphs (KGs), which can shed insights on how debunking can reduce information distortions. Also, studying differences in source credibility of debunking messages on WhatsApp is a shift from the existing approaches, which are either experimental or focused on social media like Twitter.