Cyberbullying and social media user-verification
Social media platforms enable people to communicate in both positive and negative ways, including in ways that may be abusive and bullying. Abusive messaging can harm mental health, and has been shown to increase during periods of public crisis, such the Covid pandemic. There is a need to better identify and classify cyberbullying and online abuse, to improve the design of deterrence strategies.
In a recently published study VISION researcher Dr Lifang Li explored how the ‘verification status’ of social media user accounts was associated with cyberbullying. Verification refers to when a social media user’s identity has been confirmed, for example by the checking of an identity card. Lifang examined data from China’s main social media platform, Weibo, to classify messages that had been posted during the pandemic about people who were diagnosed with the coronavirus. She examined the content of posts made by users who were verified and unverified, used techniques to understand how often anger-related words were used, and measured the extent to which the posts got shared.
Posts that could be classified as critical of people diagnosed with Covid during the pandemic (for example, describing them as ‘reckless’ or ‘selfish’ for having contracted the infection) were in the minority, most social media users were understanding or neutral in their online communications. Lifang found that posts that were critical of people diagnosed with Covid were more likely to use anger-related words. Although not a focus of the paper, official verification of a social media user’s identity did not appear to be strongly related to how likely they were to post or repost critical views.
However, male verified social media users were more likely than unverified or female users to have their posts shared. This suggests that their online activity may have a disproportionate impact on other users. Cyberbullying monitoring may need to consider such differences, especially in the context of public health crises.
This study made novel use of machine learning techniques, which may help other researchers developing algorithms to identify abusive posts online.
For further information, please read the publication at Frontiers | Social media users’ attitudes toward cyberbullying during the COVID-19 pandemic: associations with gender and verification status (frontiersin.org) or contact VISION researcher and study co-author Angus Roberts at angus.roberts@kcl.ac.uk.
Photo from licensed Adobe Stock library
