Measuring Behavioral Gender Homophily in Co-Authorship Formation

• Activity Number: 410 - Social Issues, Trends, Inequality, and Employment
• Type: Contributed
• Date/Time: Tuesday, August 1, 2017 : 2:00 PM to 3:50 PM
• Sponsor: Social Statistics Section
• Author(s): Y. Samuel Wang* and Elena Erosheva
• Companies: Dept of Statistics and University of Washington
• Keywords: homophily ; gender bias ; social networks ; assortativity ; co-authorship

Abstract

Collaboration and co-authorship can be critical to a researcher's scientific and career advancement. In this work, we examine the role of gender in co-authorship formation and test whether a propensity to co-author with others of the same gender, which we refer to as behavioral gender homophily, can be detected in a large-scale analysis of the JSTOR corpus of scientific publications. Much of the previous work has analyzed select journals in a single discipline -- most prominently, economics -- showing that substantial gender homophily exists in co-authorship. However, these findings could be confounded if researchers collaborate primarily within sub-disciplines -- tight scientific communities that compose a discipline -- and gender ratios and co-authorship norms differ substantially across sub-disciplines. In our analysis, we adjust for differing gender ratios and co-authorship norms across sub-disciplines and use a MCMC approach to perform a non-parametric test for behavioral gender homophily. This presentation will focus primarily on the statistical aspects of the project.