Emergence of echo chambers in social media communication involving multiple topics

In recent years, research on the existence and impact of echo chambers has proposed mixed perspectives. Several studies have pointed out that actual social media networks are segregated based on users' preferences regarding specific topics, limiting opportunities for exposure to opposing opinions and thus suggesting the formation of echo chambers. On the other hand, several studies suggest that echo chambers are overstated, as the followers of mainstream media provided opportunities to be exposed to information that contradicts their own preferences. We assumed that these conflicting perspectives arise from variations in the diversity of topics in their research area. Therefore, we examined how the increase in topics within communication spaces on social media affects the formation of echo chambers. Experiments based on Multi-Agent Simulations reveal that echo chambers were formed when the number of topics was small, but as the number of topics increased, the formation of echo chambers was inhibited. However, in situations involving multiple topics, it is improbable that individuals will address all topics equally. To address this, we applied the concept of agenda-setting effects, enabling agents to recognize topics with more exposure as important and adding conditions to assign weights to those topics in additional experiments. As a result, the network of agents became echo chambers regarding topics recognized as important by most agents. From these results, our study suggests that the contradictory perspectives regarding the existence of echo chambers in related research can be reconciled by considering the 'number of topics' and 'the weights of topics.'