This study examines the team-level effects of pair programming by developing a research model that accounts for mediators and moderators of the relationship between pair programming and team performance. We hypothesize that pair programming helps software development teams establish backup behavior by strengthening the shared mental models among developers. In turn, backup behavior attenuates the negative effect of task novelty on team performance. We collect data from the software developers, Scrum masters, and product owners of 62 software development teams in a global enterprise software firm and find broad support for our research model. The study makes important contributions by shifting attention to the team-level effects of pair programming and by explicating mediating and moderating mechanisms related to the roles of shared mental models, backup behavior, and task novelty. The results underline the importance of viewing pair programming as a context-specific practice that helps establish backup behavior in teams. In terms of implications for practitioners, our results show that pair programming can be a valuable element of team governance to create shared mental models and backup behavior and to achieve high team performance when teams face high levels of task novelty. Link to the article
KUDE, T., MITHAS, S., SCHMIDT, C.T. and HEINZL, A. (2019). How Pair Programming Influences Team Performance: The Role of Backup Behavior, Shared Mental Models, and Task Novelty. Information Systems Research, 30(4), pp. 1145-1163.