Hashtag homophily in twitter network: Examining a controversial cause-related marketing campaign

Social media such as Twitter generate vibrant discussions related to key sociopolitical issues and have great ability to project various discourses into public arena. Yet, these discourses can be overwhelming and heated, in particular when controversial events happen. In this study, we performed topic modeling on more than 100,000 original tweets to examine Twitter discourses around Gillette's controversial cause-related marketing in 2019 and conducted network analysis through exponential random graph models (ERGM) to investigate the homophily tendency of users who used certain hashtags. Results show that users' discourses were mostly within the message frameworks of the campaign but users strongly reacted to others and top influencers in their online discussions. In addition, the mention network of these users showed homophily tendency based on hashtags. Homophily in this study was distinguished based on attraction of common users (i.e. increased chance of ties for users who both engage the hashtag) and alienation of nonusers (i.e. decreased chance of ties for users neither of whom engages the hashtag), and the comparison between ideological hashtags and conceptual hashtags revealed that homophily only manifested through ideological hashtags.