Toxic content and social media user engagement

New research finds that toxic social media content increases user engagement but cautions that engagement and welfare are not necessarily aligned.

Social media platforms play a balancing act between the drive to increase user engagement on their channels and reducing exposure to toxic content.

In the US, more than seven in ten people are active on social media and most users report experiencing some form of online harassment in their lifetime.

New evidence Toxic Content and User Engagement on Social Media from George Beknazar-Yuzbashev, Rafael Jiménez-Durán, Jesse McCrosky and Mateusz Stalinski draws on existing links between inflammatory content and violence, mental health and politics.

Social media algorithms, which are trained to maximise various forms of engagement may be inadvertently amplifying toxic material. Work in social psychology has highlighted these concerns by showing that negativity can have a particularly engaging impact on human behaviour.

The issue of toxic content has been under public scrutiny in recent years but there has been a scarcity of evidence on the links between toxic content and user engagement.

Using a field experiment with 750 social media users who installed a browser extension able to automatically hide toxic content, the research investigated the impact of toxic content on user engagement across three major social media platforms—Facebook, Twitter (now X), and YouTube.

A survey experiment of over 4000 participants further examined curiosity-driven engagement and welfare effects of exposure to toxic content.

The results of the field experiment reveal that lowering exposure to toxic content reduces user engagement across various measures. In particular, users spend less time on social media and have fewer ad impressions and clicks, leading to a potential revenue loss for platforms. Moreover, hiding toxic content causes users to substitute to other platforms, where toxicity is present.

The survey experiment shows that while users are more likely to engage with toxic content, they do not necessarily enjoy engaging with it.

Mateusz Stalinski, Assistant Professor of Economics and CAGE Research Associate said:

“Our experiments found that social media platforms face a trade-off between curbing toxicity to reduce harm, and the lowering of engagement and advertising revenue.

“It highlights that platform incentives do not align with social well-being, as engagement-driven algorithms may amplify toxic content.

“Policy interventions are needed to ensure social media companies balance content moderation with financial incentives.”

• Read the full report: Toxic Content and User Engagement on Social Media: Evidence from a Field Experiment