|
 |
Web services allow users to send messages to each other. Thus, exploring the structural features of human networks that contribute to the increased activity of a service is both an academic question for researchers and a management interest for companies. However, some indices reflecting the characteristics of the network structure show the characteristics of individuals in the network, while others show those of the entire group (network). The analyst must understand these differences in characteristics before measuring the influence of indices on human activity. In addition, network indices are interrelated, and it is difficult to determine which index has the true causal effect. In this study, we select indices and apply a method that combines multilevel SEM (ML-SEM) and a cross-lagged model (CLM). Using Unipos data with network structures, we search for indices that affect monthly post frequency from among centrality and a scale-free index calculated based on an individual and a group, respectively. The results of our analysis show that closeness centrality has consistent and significant positive effects on post frequency in subsequent periods.
Research papers (proceedings of international meetings)
