Social Network Sites (SNS), the most prevalent Web 2.0 service, blossom with interpersonal sharing practices. A culture of sharing and the subsequent production of abundant social information are driven by the visibility of digital networked-ness. Social information embedded in digital social networks influences the shaping of our attitudes, thoughts, and online behaviors. Marketers and campaigners have been taking advantage of SNS's social information for various instrumental goals. Based on a cyber-field behavioral experiment on Facebook, this dissertation attempts to theorize underlying mechanisms of the SNS social influence process and analyze the ensuing formation of a collective communication structure. This dissertation particularly emphasizes the social network effects on the influence outcome. Multilevel-perspectives are employed to analyze the dynamics of Facebook's social network influence.

The findings are summarized as follows. First, an individual-level examination was performed. Particularly, personal influence on mobilizing others' online behaviors was explored. The results found that the influentials in SNS were characterized as not only having a leader-like personality, but also being digitally connected to a diverse and large number of social contacts through participation in multiple virtual group activities and maintenance of heterogeneous personal networks. The findings of this project support the concept that the personal influence on Facebook should be understood as normative rather than informational.

Second, structural social influence has been theorized based on three sub-mechanisms: Direct recommendation, social contagion, and network embeddedness. The study found the following: (a) message compliance was stronger when individuals were exposed to multiple direct contacts; (b) Facebook displayed the effect of indirect exposure to others' behaviors, not just the effect of direct recommendation; in other words, it has a social contagion effect; and (c) While network embeddedness did not directly affect an individual's online behavior, this positional property acted as a moderator for direct recommendation and social contagion effects.

Lastly, a macro-structural analysis was conducted to explore the properties of communication systems that emerged through social influence processing. The community structure strategically formed by this field experiment was examined based on the three well-known network topologies—scale-free, small-world, and centralization, finding the emerged community was characterized as scale-free and small-world like and weakly centralized. The implications of having such structural properties on the effectiveness of communication system are discussed.