Expertise and Complexity as Mediators of Knowledge Loss in Open Source Software Development
Olivia B. Newton, Stephen M. Fiore, Jihye Song · Proceedings of the Human Factors and Ergonomics Society Annual Meeting · 2019
This paper describes an approach integrating cognitive engineering with computational social science in the context of open source software (OSS) development. Through an analysis of large-scale collaborations in a complex operational setting, we study how expertise and task complexity predict changes in productivity when knowledge loss occurs. Using team data from thousands of software files, we model the effects of expertise, complexity, and knowledge loss on productivity. On its own, knowledge loss had a negative effect on productivity, but this effect was reversed when knowledge loss was combined with high complexity and high numbers of newcomers, suggesting that experts are better able to utilize crowdsourced work. We identify opportunities for research to inform prediction of outcomes in OSS projects based on team and task characteristics and demonstrate the value of integrating cognitive engineering with computational social science to study collaborative work in sociotechnical systems.