A Closer Look into Collaborative Publishing at Software-Engineering Conferences

Computer science and particularly software engineering is a rapidly evolving research discipline increasingly conducted by large, collaborative teams. Unfortunately, there is little research on the underlying publication activity and collaboration patterns in software engineering. To address this gap, we study two properties of research collaborations in software engineering: the number of collaborators (i.e., authors of a paper) and their academic age (i.e., their experience of working in research). More precisely, we investigate collaborations for papers published at all main tracks of three top-level software-engineering conferences (i.e., ASE, ESEC/FSE, ICSE) and one top-level reference conference (i.e., JCDL), including a total of 5,188 papers and the corresponding 8,730 unique authors. Our results indicate that collaboration is more prevalent now than ever before, with a decline in the proportion of researchers who contribute single-author papers. Moreover, our analysis revealed that the ideal team size seems to range from two to four researchers, and that junior researchers seem to need the support of more experienced co-authors to get published at such top-level conferences. Ultimately, our goal is to understand how collaborations in software engineering have evolved and impact different researchers (e.g., newcomers, juniors), helping to highlight potential impediments and consequent improvements regarding the quality of research, collaborations, and mentoring.