| As more of our society participates online to perform everyday activities from shopping to socializing, the traditional methods of diffusion are changing. These changes are important, not only for individuals participating, but also for the organizations and businesses that are entering the online space. Information of new innovations and ideas spread through societies by means of interaction and communication (Rogers 1976). Understanding how diffusion functions online gives all the stake holders information that can be used to optimize their online experiences. Historically, diffusion followed orderly paths dictated by networks constructed from relationships of who had contact with whom (Bala and Goyal 1998). Today, the ability to reach others has shortcut the traditional paths. Creating communities comprised of relationship possibilities is only bounded by technologies. Diffusion in today's communities requires an understanding of the member behaviors that lead to the formation of the relationships that become the pathways enabling exchange. Based on fundamental theories about social networks and diffusion, this research proposes a Member Diffusion Behavior Model (MDB) that explains member diffusion behavior with three antecedents: Informal Sharing (IS), Promotion (Pr), and Persistent Conversation (PC). The findings of this research toile in three areas, the testing of the behavior model, the discovery of the behaviors from conversation structure, and the identification of the influence of seed members on diffusion. Hypotheses testing provides support for the IS and PC paths in the MDB model. Using a new structure based method, these two behaviors IS and PC are shown to be successfully discovered from the conversation structure. When the top defusing members identified using the structure metrics are used to seed the diffusion process, these members are found to be more effective than members selected using the traditional degree centrality method. This research contributes to the understanding of diffusion in social networks in several important ways. First, it provides a theoretical model to explain member diffusion behavior by integrating theories of diffusion and social networks. Second, it designs methods and develops metrics for measuring diffusion behavior in social networks by integrating social network analysis and information retrieval techniques. Third, it helps to gain insight into the diffusion process by exploring how members with influencing behaviors affect the network. |
