How Pair Programming Influences Team Performance: The Role of Backup Behavior, Shared Mental Models, and Task Novelty

Many organizations and software firms these days use pair programming to meet the challenges posed by digital transformations and to meet customer needs. They use pair programming to become more agile and responsive to their competitive environment. Although some studies suggest that pair programming can potentially increase the quality of developed software code and the satisfaction and confidence of developers, very little is known as to how pair programming affects team performance. Therefore, we conducted an empirical study at one of the largest enterprise software firms that serves a large number of Fortune 500 companies, and we collected data from the software developers, Scrum masters, and product owners of 62 software development teams. Our findings show that pair programming affects team performance positively by changing cognitive structures and behavioral patterns in software teams. Pair programming helps team members develop shared mental models and, as a consequence, increases backup behavior among team members. Backup behavior is particularly valuable for teams facing high task novelty. Our findings have critical implications for organizations and senior managers as they lead their digital transformation efforts where they often rely on autonomous teams for providing digital innovation. We show that pair programming can be a key ingredient for high-performance teams. Establishing the practice of pair programming promises to be particularly valuable for those teams that lack shared mental models and teams whose members fail to provide each other backup.

• 저널Information Systems Research · 30(4) · 1145–1163
• 토픽Software Engineering Techniques and Practices · Information Systems
• DOI10.1287/isre.2019.0856
• 저자Thomas Kude, Sunil Mithas, Christoph Schmidt, Armin Heinzl