An Empirical Study of Social Debt in Open-Source Projects: Social Drivers and the “Known Devil” Community Smell

Social debt, the accumulation of unforeseen project costs from suboptimal human-centered software development processes, is an important dimension of technical debt that cannot be ignored. Recent research on social debt focusing on the detection of specific debt indicators, called community smells, has largely been conceptual and few of them are operationalizable. In addition, studies on the causes of community smells focused on group processes instead of individual tendencies. In this paper, we define and investigate four social drivers-factors that influence individual developer choices in their collaboration-in 13 open-source projects over four years: 1) inertia, 2) co-authorship (by chance or by choice), 3) experience heterophily, and 4) organization homophily. Building on previous studies and theories from sociology and psychology, we hypothesize how these drivers influence software quality outcomes. Our network analysis results include a contradiction to existing studies about experience heterophily and reveal a new community smell, which we call “Known Devil”, that can be automatically detected.