Assessing individual performance in Agile undergraduate software engineering teams

The Agile Software Development (ASD) process is at the forefront of rapid product development driven by changing customer requirements and a trusted, self-organizing development team. Scrum has become a viable model of ASD focusing on determining immediate deliverables and structuring short timelines, called Sprints, for designing, implementing, and providing them for testing by the customer. While these practices are being adopted by organizations, there is significant difficulty in scaling them to the classroom. Once in place, it is a complex task to evaluate individual student performance based solely on the product outcome and Sprint grade. Thus, there is limited opportunity to catch performance problems that may lead to missing deliverable deadlines or decreasing team trust. In this paper, we impose ASD using Scrum on a senior software projects course in Computer Science. Using a collaborative environment that embeds a social network, project management modules, and event capture system, we perform broad data and event capture and analysis to investigate metrics that are relevant to assessing individual performance aspects related to functioning on an Agile team for software development. Our results suggest that predictive data is available after each Sprint to ascertain individual performance attributes and their relationship to product outcomes.