An Analysis of Human Aspects of Collaborative Group Members in OSS development

Abstract

Open Source Software development is a collective activity that involves different software developers who may differ from each other. Although, previous researchers have focused on technical aspects like code factors, technology used etc, recently researchers have explored non-technical human aspects like personality, ethnicity and gender to measure various outcomes. This research assists the emerging state-of-the-art body on diversity research with an empirical study that analyzes how the personality and, the race and ethnic diversity of members in a collaborative group relates to their collaborative contributions in Open Source Software(OSS) development. This research contains two parts - In the first part we analyse the collaborative group members’ personalities and frequency of their collaborative contributions. In the second part we analyse the relationship between the diversity of collaborative group members’ race and ethnicity, and the frequency of their collaborative contributions in GitHub. We infer collaborative groups within a project based on the collaboration between software developers in that project. Since previous studies have shown pull requests as the major contribution for a developer to be accepted as a group member, we measure the collabora- tive contributions of the group members by the number of pull request the group members have merged collaboratively. Our results from the first part of our research, indicate that 1) the personality traits of collaborative group members does have a relationship with the frequency of their collab- orative contributions. Specifically, the more conscientious and less extroverted the group members are, the more contributions that the group members merged. Furthermore, 2) groups that are more diverse with respect to Conscientiousness or Neuroticism have a negative relationship with the frequency of their collaborative contributions. Finally, 3) collaborative groups that are having a majority of highly open, conscientious, or neurotic developers have a positive relationship with their collaborative contributions as well. Also, from the second part of our research, We observe that (1) a major part of the developer population are White developers; (2) homogeneous and heterogeneous collabo- rative groups, with respect to race and ethnicity of their group members, have a different distribution of collaborative contributions, with heterogeneous groups having more num- ber of contributions than homogeneous groups and (3) Diversity of race and ethnicity of members in a collaborative group does have a positive statistically significant relationship with the frequency of collaborative group members’ contributions.