Misinformation and Conformity in Online Group-Based Interactions

• In Event: Experimental Research on Partisan Polarisation and Misinformation

Abstract

Misinformation is now widely acknowledged as a leading threat to democratic politics, especially since social media offers a fertile ground for its proliferation on a larger scale. Over the past decade, scholars and policymakers have sought to better understand the psychological mechanisms leading to belief in online misperceptions, rumors, and conspiracies. Most of the research on the determinants of belief in online misinformation has so far focused on exposure to misinformation through what may be referred to as feed-based interactions – that is, online interactions that take place on interfaces to which access is not restricted to a specific subgroup. Yet many online interactions today happen within private groups, on enormously popular platforms such as WhatsApp, Telegram, and WeChat, among others. These group-based interactions take place more privately, among self-organized communities defining themselves as separate, exclusive groups assuming a specific group identity.

Over the past few years, such group-based interactions have been at the heart of numerous controversies across developed and developing countries. News reports have repeatedly suggested that group-based interactions distort beliefs among the electorate and beyond, which may in turn increase individuals' propensity to engage in hostile, radical, or violent behaviors. Yet, despite such alarming reports, very little is known about the mechanisms through which these interactions may contribute to the global misinformation crisis. How, if at all, might online group-based interactions facilitate belief in misinformation?

In this paper, we argue that key affordances of group-based platforms can increase belief in misinformation. Our argument takes the group nature of interactions seriously and highlights the identity dimension involved in interactions on such platforms. Because of their group-based nature, and because of the existence (by design) of ties between respondents, we argue that group-based interactions may foster a common identity between their participants, which may in turn lead to increased trust.

We use an innovative WhatsApp-based field experiment in India to investigate the dynamics of group-based interactions and belief in misinformation. Respondents are randomly assigned to WhatsApp group threads where they interact with other respondents as well as confederates from the research team. We introduce various claims and headlines in each group thread, varying the degree of ingroup endorsement of each claim and the degree to which group participants share social identities. This design allows us to evaluate how the perceived degree of unanimity and homogeneity in a group can shape beliefs in various true and false claims introduced by group members, and whether these characteristics in turn affect polarization and other downstream attitudes.

This research contributes to several streams of research related to misinformation. First, we focus on the case of an understudied medium where much of the online (mis)information in the global South circulates. Second, as we do so, we take into account the specific affordances of this medium and the social-psychological dynamics it may trigger. Third, we build for the first time an on-platform field experiment that allows us to test the effect of features of these groups in a relatively naturalistic setting.

Authors

• Simon Chauchard, University Carlos 3 of Madrid

• Rajeshwari Majumdar, Yale University