The echo chamber effect on social media has attracted attention due to its potentially disruptive consequences on society. This study presents a framework to evaluate the impact of personality traits on the formation of echo chambers. Using Weibo and Twitter as platforms, we first define an echo chamber as a network where users interact solely with those sharing their opinions, and quantify echo chamber effects through selective exposure and homophily. We then employ an unsupervised personality recognition method to assign a personality model to each user, and compare the distribution differences of echo chambers and personality traits across platforms and topics. Our findings show that, although user personality trait models exhibit similar distributions between topics, differences exist between platforms. Among 243 personality model combinations, over 20% of Weibo echo chamber members are "ynynn" models, while over 15% of Twitter echo chamber members are "nnnny" models. This indicates significant differences in personality traits among echo chamber members between platforms. Specific personality traits attract like-minded individuals to engage in discussions on particular topics, ultimately forming homogeneous communities. These insights are valuable for developing targeted management strategies to prevent the spread of fake news or rumors.
Read full abstract