From Strangers to 'Friends' Online: A Multi-Theoretical, Multi-Level Investigation of the Structure of Social Relationships in Virtual Worlds

This dissertation explores the nature of online social relationships, examining the relative influence of individual, dyadic and network-level pressures on the emergence of social relationship ties on the Internet. Using Monge and Contractor's (2003) multi-theoretical, multi-level network modeling and data drawn from logs of social behavior inside the virtual world Second Life, this research adds a much-needed network analytic perspective to existing literature on online social behavior. The results of predictive network analyses reveal that usage behavior and the network pressure towards balance/triadic closure best predict the emergence of social relationships across all sampled data. Comparing networks of adolescents and adults separately reveals that adolescents are particularly susceptible to network balance influences, while adults are more likely to form relationships based on age homophily. These results are contextualized with the results of qualitative interviews with Second Life residents, which offer insight into why the observed trends emerged. Results are discussed in light of existing research on online social relationships, and a new theoretical model for predicting social relationship formation is proposed.