Eric is a computer scientist in the Computer Science Laboratory at PARC. Since receiving a B.S. from Caltech in 1980, he has spent time as a video game designer, developing high-speed imaging software, and creating research software prototypes, as well as living and doing language- and culture-learning for several years in an isolated village in Southeast Asia. He also holds an M.A. from Fuller Theological Seminary. As well as an interest in developing and using tools to understand the social nature of online role-playing games, long-term interests include algorithms, and the processes and tools by which groups of people develop high-quality software and enjoy doing it.
Leveraging Virtual Omniscience: Mixed Methodologies
for Studying Social Life in Persistent Online Worlds
Massively multiplayer online worlds (e.g., EverQuest or SecondLife) constitute a new form of social life that is ripe for social scientific study. Inhabitants of persistent virtual environments spend on average more than twenty hours per week logged on. They interact with each other through a simulated face-to-face, they share an emergent in-world culture, they form long-term relationships, and in some cases they even make a real-life living from creating and trading virtual property. One thing that makes research on such virtual worlds particularly exciting is the fact that, in principle, near-total data on all player activities in-world can be logged by the system. In other words, a kind of virtual omniscience is technically possible. Leveraging this kind of virtual-world data presents new analytic challenges and requires a mix of social science methods. The PlayOn project at PARC is tackling these challenges by using a three-pronged methodology. First, we use participant observation to gain a member's perspective of the game world. We spend many late nights playing the game, mastering game play, learning game culture, and joining player associations. Second, we use conversation analysis to understand the micro-level organization of the simulated face-to-face interaction. Using real-time video screen capture and annotated chat logs of player interaction, we specify the formal features of avatar-and-chat-based interaction and compare them to those of real-life face-to-face. Third, we use social network analysis to measure the macro-level organization of player-to-player interaction, which entails data mining player-activity logs for group-level interactivity patterns. We build tools for parsing activity logs, computing social interactivity metrics, and visualizing social network and traffic patterns. The participant observation in turn informs both the social network analysis and conversation analysis.
In the first third of this 90-minute workshop, we will report on our study of Star Wars Galaxies with a focus on the challenges we faced in collecting and analyzing the data. In the second third, we will open discussion of the methodological challenges in studying massively multiplayer online worlds and encourage the audience to share their methodological approaches and experiences. Finally, in the last third of the session, we will take a second look at our raw data from Star Wars Galaxies and invite the audience to explore it again with us.