Abstract
A growing body of research is using artefacts from online development communities to explore the impact of developers’ behaviours on the software development process. Although this research has produced many insights, researchers have yet to fully explore the impact of developers’ cultural backgrounds on their behaviours in an online community, although such understandings could be useful for helping the community to understand and plan for team dynamics. This study utilised a pragmatic case study to explore the relationship between culture and online behaviour among developers from the United States (U.S.), China, and Russia—three countries that differ in their orientations as individualistic or collectivist cultures. The data for the study comprised artefacts supplied over an 11-year period by users of Stack Overflow1, a popular online programming community that addresses questions from members by providing them with rapid access to the knowledge and expertise of their peers. Artefacts consisted of developers’ questions and answers, personal profiles, Up and Down voting records, online reputations, and earned badges. Data mining techniques, as well as statistical, linguistic, and content analysis were used to compare artefacts from the three groups of developers based on their cultural orientation as individualistic or collectivistic, attitudes, and interaction and knowledge sharing patterns. The findings revealed differences among the three groups that were consistent with their cultural backgrounds. U.S. developers, who are from an individualistic culture, asked and responded to more questions, had higher average reputations, used the pronoun “I” more frequently, and were more task- focused. Conversely, Chinese developers, who are from a collectivistic culture, provided more extensive commenting and editing of posts, used the pronouns “we” and “you” more frequently, and were more likely to engage in information exchange. Russian developers had been using Stack Overflow the longest and were the most reflective. The cultural patterns identified in this study have numerous implications for enhancing in- group interactions and behaviour management among software development communities.