Findings

This chapter presents the findings from the three empirical investigations conducted in this study. An initial exploration yields insights into the structure of collaborative quality assurance techniques and shows that different teams use the same techniques for distinct tasks and towards various ends. That is, on a task level, the techniques provide distinct functional affordances to different teams. Subsequently, an in-depth analysis of four ISD teams and their collaborative quality assurance practices provides a detailed understanding why different teams perceive distinct functional affordances. Qualitative differences in teams’ transactive memory systems can be seen as a major driver of these perceptions. The in-depth case studies also yield insights into the emergent effects of continuously applying collaborative quality assurance techniques on a team level. While pair programming fosters knowledge transfer and creates redundant knowledge in a team, peer code review fosters the task-based retrieval and application of specialist knowledge by supporting transactive memory processes. A questionnaire-based survey of 81 ISD teams with more than 600 individual respondents corroborates these findings and relates them to software quality as an outcome variable. Results show that frequent and team-wide peer code review can partially substitute for extensive transactive memory structures in producing high-quality software. In teams with distant personal relationships, pair programming can help team members build the common ground to access one another’s expertise while teams with already close relationships can harm their transactive memory with too much pair programming. Peer code review helps directly improve software quality and reduces the dependence of software quality on existing transactive memory structures.