Learning By Migrating: A Computational Study Of Diversity And Team-Level Decision-Making

How does previous experience and learning influence a team's ability to successfully agree on a system architecture, team roles and responsibilities, and design method? Migration of team members leads to diversity in past experiences and beliefs, which might have a positive or negative affect on team decision-making. Using computational modeling of self-managed teams across multiple project life cycles, we perform controlled experiments to evaluate performance and decision-making patterns of migrating vs. non-migrating teams. We find that there is no difference in mean performance, indicating that neither approach is intrinsically better. However, statistical tests of paired trials shows a meaningful an advantage for migrating (diverse) teams. Examining patterns of decision-making over time reveal that migrating (diverse) teams explore a wider range of team-level decisions, which makes them more adaptable in specific circumstances.