Reaching To The Top: The Gender Effect In Highly-Ranked Academics In Computer Science

Science, Technology, Engineering and Mathematics (STEM) disciplines are notorious for gender imbalance favouring men. From these, Computer Science appears to among the worst. The imbalance comes despite the fact that, the proportion of women publishing impactful papers has increased over the last couple of decades. Here, we explore the potential gender differences in collaboration dynamics of leading academics in Computer Science and contrast it with other researchers in the same discipline. This work aims to demonstrate that productivity patterns differ between men and women; understanding such patterns is fundamental to improving the current situation by driving well-thought interventions. We start by looking at topological features of academics ego-networks and their variability over their career lifespan; we found that women have a higher rate of repeated collaborators, while men have slightly more new collaborators. Second, we show gender differences in the acquisition patterns and maintenance of collaborations over the researchers’ careers. Third, we look at alters’ gender composition and gender role in introducing new collaborations to researchers and found that women are more homophilic than men and that the gender of the person being introduced (alter) tends to match the gender of the ego (the researcher). Last, we show that women reach their highest productivity levels earlier in their careers than men, but the levels attained are not at the same magnitude as men. Our work highlights the importance of studying gender differences by productivity levels, the role of gender-mixed groups in academia and the relation to academic production within these groups. This work contributes to a better understanding of gender roles in Computer Science and points to ways that can help reduce possible gaps.