Exacerbating or Attenuating? Moderating False and Provocative Claims

• In Event: Understanding and Addressing Online Political Incivility, Hate, and Polarization

Abstract

The purpose of content moderation on social media platforms such as Twitter and Facebook is to facilitate open and safe public conversations, which is commonly achieved by flagging or removing provocative, disputed and other potential harmful content. Such interventions can help to fight online extremism and false beliefs. However, by imposing limits on acceptable behavior online, content moderation may also decrease trust in public official or media outlets, and increase skepticism in otherwise uncontested pieces of information. This project seeks to address the potential repercussions of online content moderation by inviting U.S. citizens to participate in an online discussion experiment through our newly developed Tool for Online Discussion (TOD). Uniquely, TOD enables participants to engage with each other in real-time, which creates the illusion of a realistic online discussion about a given topic. In an experimental set-up, we pre-program ''bots'' that appear as one of the participants to send messages that contain either false or provocative claims. As a treatment, these claims are being flagged as either 'disputed' or 'sensitive', respectively. In the control group, these pre-programmed messages are not moderated. We measure the effect of flagging false and provocative content on participants' level of trust, their endorsement of false, factual and provocative claims, their general attitudes towards content moderation, as well as their opinion on the discussion's topic. This project sheds new light on both the effectiveness and social externalities of online content moderation, both of which to date are only poorly understood.

Authors

• Emma Hoes, Zurich University

• K. Jonathan Klüser, University of Zürich

• Nahema Marchal, University of Zurich