Predicting Team Performance through Leadership Behavior

This study provides a methodology to identify the characteristics of emergent leadership in a performing team. It is based on an embedded case study analyzing seven project teams of a distributed learning community, whose members never met face-to-face, and were organized in virtual teams collaborating on common projects. The research setting was a course taught to Information Systems, MBA and PhD students coming from geographically distant universities. We observed the communication behavior of 32 community’s members over four months by gathering their e-mail exchanges using a proxy e-mailbox for each project team. The analysis was conducted on two dimensions: team performance and leadership characteristics. “Performance” was based on peer evaluation and the number of papers completed, presented at conferences or published in journals; “emergent leadership” was identified by looking at the evolution of actors’ betweenness centrality over time. The results show that low and high performing teams are characterized by different patterns of the leaders’ behavior. The high performing teams are the ones with a member having a constantly high betweenness centrality (Cb) that decreases as the followers’ centrality increases. The low performing ones are those with a member having a constantly high Cb but with his/her Cb increasing as the followers’ Cb decreases. Finally, the lowest performing teams are characterized by a form of fluctuating leadership, where no member holds a leadership role. The study provides empirical evidence on how to recognize a successful project team: 1) the presence of the same individual in a leadership position over the entire community life cycle; 2) a decreasing value of the leader’s betweenness centrality as the followers’ betweenness centrality increases.