New emergent forms of online communication started to change teaching and learning processes. It brought about online learning that has gained more and more attention from educator, learners and researchers. However, although communication is considered an essential element of the learning process, little is known about how online learning environments shape students’ interactions. The current paper sheds light on the structure and properties of social networks formed in an online learning environment. Two instructional approaches were compared – one incorporating traditional lecturing and blogging, and the other including traditional lecturing, blogging, and a Multi-User Virtual Environment (MUVE) platform called Second Life (SL). 56 students in a large Mid-Wester university enrolled in an undergraduate-level psychology course participated in the study. In the control condition, the students received a lecture and were required to write a blog post every week. In the experimental condition, the students received a lecture, wrote a blog post each week and also participated in SL activities during class time. Students’ interactions on the blog in both conditions were coded and analyzed at two time points: at Post 1 and Post 12. Social network analysis was used to analyze the network density, centralization and to calculate individual node centrality indices. ANCOVA and Quade’s test (non-parametric ANCOVA alternative) were used to analyze network connectedness. The results revealed the following: (1) Students in the experimental condition made more connections with each other than in the control condition, potentially due to the more egalitarian nature of communication channels in SL and more opportunities for synchronous communication; (2) The network in the experimental condition was more distributed and less centralized. Peer network analysis showed that students in the experimental condition took more control over the information flow and dispersion. The instructor played a less significant role in experimental condition network. Second Life allowed for creating an autonomous, distributed learning environment that could form a more distributed interaction network; (3) The instructor’s influence in both class networks decreased over time; however, he remained a facilitator of the information flow in both conditions, although to a much lesser extent in the experimental condition as compared to the control condition. The implications of these findings, possible limitations and future research directions are also discussed.