Brian C. Keegan, Ph.D. - US grants

This research leverages the large-scale, detailed, international, and unobtrusive data logged by online electronic sports (e-sports) to understand the organizational behavior of temporary virtual teams. Teamwork and collaboration are essential for success in contemporary organizations, and as teams become increasingly distributed, virtual, self-assembled, cross-functional, and temporary, existing frameworks for supporting effective teamwork need to be revised. Multiplayer online e-sports are models for developing new and more effective frameworks for technologically-mediated teamwork: clear and consistent performance metrics, detailed and public behavioral data for quantitative analysis, a large and international user base, and extensive randomization, repeated observations, and matching for making strong causal inferences. The inferences made from detailed behavioral records of high-tempo, naturalistic decision making can be extended to many other settings such as disaster response or breaking news. The findings from this research could provide the empirical basis for identifying under-utilized expertise in noisy social systems, optimizing team assembly algorithms to improve performance, improving decision making in temporary virtual teams, and using online e-sports as diagnostic tools for existing teams. The findings from this project will also inform the design and governance of e-sports that already attract tens of millions of users around the world.

The project pursues two initiatives to understand how to improve the performance of temporary virtual teams. The first initiative examines how team assembly decision-making in high-tempo contexts influences team performance. The second initiative examines how software and database patches disrupt mental models and decision-making. This research will triangulate between (1) existing organizational theories and constructs about team processes, (2) quantitative methods from data mining, machine learning, and econometrics for analyzing unobtrusive observational data about user behavior, and (3) the unique affordances of several popular e-sports to examine the variables and mechanisms that influence team performance in a naturalistic setting. The results of this research project will (1) provide comparative empirical insights into a rapidly growing cultural and economic phenomenon; (2) develop frameworks and models to increase engagement in sociotechnical systems; and (3) provide generalizable recommendations for improving the performance of temporary virtual teams. This project will create free software libraries, data collections, and supporting tutorials and documentation enabling other researchers to develop and evaluate organizational theories using data drawn from e-sports.

This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

The project will study the emergence of online groups by examining their overlooked but important connections to existing groups. New online groups do not appear spontaneously, but are created by users who participated in existing online groups. A new group's early members carry their own activity history, which reveals their past group memberships as well as the relations between existing groups and this new group. The ability to trace the behavior of users before they start new groups presents a unique opportunity to understand how new groups emerge, how social norms arise in new groups, and what factors contribute to the success of new groups. This research will quantitatively model the genealogical relationships between online groups by tracing the sociotechnical lineage of "child" groups through their early members' previous participation in "parent" groups. Platform administrators or group moderators can use genealogical approaches for adopting existing norms in other groups, recommending new groups to users, and identifying opportunities to create new groups. A genealogical perspective can also explain the variability in group success as well as how norms spread throughout online groups. This work will enable transformative changes in online communities, either via redesign of their platforms or through implementation of new community policies.

The project draws on theories about formation of online communities, organizational ecology, and kinship to explore how a group's position within a genealogical graph influences the group's identity, norms, and success. This research will analyze log data about user behavior over time from Reddit and Wikipedia. Aggregating the socio-technical lineages of multiple groups together generates a genealogical graph documenting how new groups emerge from the old. These genealogies will be evaluated across platforms, users, and time. The project will advance human-centered data science methods by employing qualitative methods, such as interviews and focus groups, to validate the proposed genealogical graphs, which will inform quantitative methods for building genealogy graphs from large-scale log data sets and analyzing their structure and dynamics.

This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.