Recent years have witnessed the prevalence of online knowledge-sharing platforms, in which user management and content creation are the central issues to the governance. Interactions take place through users’ linguistics, which helps shape users’ perceived personalities in the eyes of others. To understand the impact of perceived personality on online discussions, a novel method that combines natural language processing method and unsupervised learning is developed to extract contributors’ perceived personalities based on the contents they generated on the platform, which is further validated by a lab experiment. An empirical analysis is then carried out to unpack the role of different personality dimensions on online discussions. Our results reveal that the first contributor’s perceived conscientiousness and openness exhibit a significant but contrary role on content generation. Our method and empirical analysis can provide insights into the governance of online knowledge-sharing platforms.

Abstract Only