摘要：This study examines how group size, network density, participation equality and member turnover are related to communication activity and shared language in online communities. It examines the social interactions of sixteen randomly selected discussion groups (33,450 participants and 632,622 messages) covering topics in politics, health, hobbies, and science and technology over a 20-month period. It relies on social network analysis, computational linguistics and hierarchical linear modeling to uncover the extent to which group attributes impact message replies, conversational thread length and shared word choices. The results show that group size and network density are positively related to communication activity and shared language, while participation equality and member turnover are not significant predictors.
