Modeling Gender Differences in Membership Change in Open Source Software Projects

Abstract: Gender diversity in open source software development continues to be a topic of growing interest among researchers, practitioners, and organizations. To date, research has revealed disparities in participation between developers on the basis of gender, with women being significantly underrepresented in open source development. Using a large data set curated for studies of diversity in open source projects, we contribute to this body of work by characterizing the relationship between gender-based participation differences and group composition in GitHub. We found that contributors identified as women and contributors of unknown gender have a shorter tenure in open source projects compared to those identified as men. Additionally, at the team level, we found that project teams with mixed-gender composition were associated with lower turnover and teams with greater disparity in the distribution of platform tenure were associated with higher turnover. Finally, our case study reveals that when looking at the entire team, GitHub users were more likely to remain in a project rather than leave after contributing but when looking at women specifically, users were more likely to leave rather than remain after contributing to a project.