Funded by the Army Research Office
Award number W911NF-14-10686.
Socio-Cognitive Networks: Theory & Data Driven Approaches for Understanding the Assembly and Interaction Networks of High Performance Teams
Military organizations are built around small units (i.e., as teams). The success of soldiers working in teams to perform tasks ranging from peacekeeping to security operations is the cornerstone of the US military. Given the importance of teamwork to military effectiveness, a key priority is developing insights into the assembly and interaction networks of high performing teams. This ARO project is a bold interdisciplinary effort to bring computational social science data collection and analytic methods to sociological theories of group self-organization and industrial psychological theories based on input-process-output models of team performance to address the elusive and practically important problem of accurately explaining team performance. As such, this project fuses heretofore-disparate theoretical perspectives with cutting edge computational methods in the interest of developing, to borrow a term, a “Moneyball” for scientific teams. We are conducting an in-depth analysis of a very rich data set on the assembly, interactions, and outcomes of scientific teams interacting in nanoHUB – a cyber-infrastructure tool supported by NSF and deployed over the past decade to a community of users that has grown to more than 250,000 annually from 172 countries worldwide. The data will enable us to study the ecosystem of teams – observing individuals’ assembly into teams, their interactions, their successes and failures, and their nucleation into new teams.
There are two main intellectual contributions from this project. First, to address the team assembly challenge, we derive from our network metrics, linguistic style matching analysis and our emotional activation analysis, a small set of high leverage factors, at the compositional, relational and ecosystem level, that can identify the assembly of virtuous combinations of teammates. Second, to address the adjustment of the team to process problems, we identify from our network metrics, linguistic style matching analysis and our emotional activation analysis, new empirical regularities that relate early warning signals of team process to team performance and regulation.
This effort is among the first to advance team theory by deploying computational approaches to measure real-time interactions in an approach that is replicable, quantifiable, and at scale. In particular, we strive to generate novel theories of team assembly, processes and performance that leverage recent advances available in cyber-infrastructure, unstructured data analytics, and social science theories.
Finally, we see pragmatic contributions. We see two pathways to the applications of these basic findings to the military. First, our understanding of the compositional, relational and ecosystem factors of team assembly can be used by the military to improve team staffing. For example, our research could be used to design team staffing systems that make recommendations based on what we are discovering are key factors in assembling high performance teams. The second pathway to application is early warning detection of team failure. The results of our research can inform the development of monitoring systems, such as the Commander’s Dashboard being developed by ARI, that would gather, process, and provide recommendations about early warning signals and interventions to military leaders.