Though research in emerging technologies is vital to fulfilling their incredible potential for educational applications, it is often fraught with analytic challenges related to large datasets. This thesis explores these challenges in researching multiuser virtual environments (MUVEs). In a MUVE, users assume a persona and traverse a virtual space often depicted as a physical world, interacting with other users and digital artifacts. As students participate in MUVE-based curricula, detailed records of their paths through the virtual world are typically collected in event logs. Although many studies have demonstrated the instructional power of MUVEs (e.g., Barab, Hay, Barnett, & Squire, 2001; Ketelhut, Dede, Clarke, Nelson, & Bowman, 2008), none have successfully quantified these student paths for analysis in the aggregate.

This thesis constructs several frameworks for conducting research involving student navigational choices in MUVEs based on a case study of data generated from the River City project. After providing a context for the research and an introduction to the River City dataset, the first part of this thesis explores the issues associated with data compression and presents a grounded theory approach (Glaser & Strauss, 1967) to the cleaning, compacting, and coding or MUVE datasets. In summary of this section, I discuss the implication of preparation choices for further analysis.

Second, two conceptually different approaches to analyzing behavioral sequences are investigated. For each approach, a theoretical context, description of possible exploratory and confirmatory methods, and illustrative examples from River City are provided. The thesis then situates these specific analytic approaches within the constellation of possible research utilizing MUVE event log data. Finally, based on the lessons of River City and the investigation of a spectrum of possible event logs, a set of design heuristics for data collection in MUVEs is constructed and a possible future for research in these environments is envisioned.