Selective matching: gender gap and network formation in research

This paper explores how the academic network extends and its consequences on research outcomes. Using a large academic bibliographic database in research in economics (RePEc), we model first the probability that two researchers collaborate, and secondly the impact of network indicators on the citation rate of research articles. Our results show the existence of a gender-based bias in the researcher matching process. Researchers are more likely to coauthor together when they are of the same gender, even when we control for productivity and proximity in the academic network before they match, as well as unobservable fixed effects of the pair of researcher. This effect is observed mostly at the beginning of the career and fades with the seniority. We also observe that network indicators have a positive impact on the citation index of research articles, suggesting that these selective matching mechanisms may have cumulative effects.