Gauging influence in software development teams

Agile software development teams are generally small, focused groups that require highly motivated members operating in a high-trust environment. Through interactive communication and collaborative work, team members can influence project outcomes both directly and indirectly. One method to examine influence is in terms of communication and control flow among team members when sharing a communication medium and artifact development tools. There currently exist several approaches for characterizing team communication using social network analysis. However, these approaches require modification to be applicable for measuring control flow influence in small teams. In this paper, we discuss the methodology and results of an influence study in which we analyze data generated by student teams in a capstone Software Engineering course. We develop three metrics to measure direct and indirect influence among team members, taking into account the temporal order of interactions and the small size of the teams studied. We correlate the influence metrics with grades, existing participation metrics, and team evaluation scores. The results suggest that our influence metrics correlate to a team member's perceived role on the team, measures of team communication and levels of task performance. This suggests that instructors can recognize the levels of influence exerted by team members by examining team communication and control flow, allowing for mediation prior to product development milestones.