Information Systems Research, 2009, Volume 20,
Issue 3, Page 377-399.
In this paper, we draw on control theory to understand the conditions under which the use of agile practices is most effective in improving software project quality. Although agile development methodologies offer the potential of improving software development outcomes, limited research has examined how project managers can structure the software development environment to maximize the benefits of agile methodology use during a project. As a result, project managers have little guidance on how to manage teams who are using agile methodologies. Arguing that the most effective control modes are those that provide teams with autonomy in determining the methods for achieving project objectives, we propose hypotheses related to the interaction between control modes, agile methodology use, and requirements change. We test the model in a field study of 862 software developers in 110 teams. The model explains substantial variance in four objective measures of project quality—bug severity, component complexity, coordinative complexity, and dynamic complexity. Results largely support our hypotheses, highlighting the interplay between project control, agile methodology use, and requirements change. The findings contribute to extant literature by integrating control theory into the growing literature on agile methodology use and by identifying specific contingencies affecting the efficacy of different control modes. We discuss the theoretical and practical implications of our results.
Keywords: agile methodologies; agility; control theory; requirements uncertainty; software development; teams