Gender inequalities in research funding: Unequal network configurations, or unequal network returns?

Despite longstanding discussions and consequent improvements of gender representation in academia, the number of women working in academic research, their performance and their recognition still indicate the persistence of gender inequalities. Of the various mechanisms that may drive such inequality the one that relates the structure of research collaborative networks to the rewards that these networks facilitate is intriguing. If social networks play a role in academic performances and women are under-recognized compared to men, what are the mechanisms that relate networks to gender inequalities in academia? Do men and women obtain equal returns from similar network structure, or do women have to build different networks to obtain similar returns? This article contributes to the theoretical debate that aims to disentangle the mechanisms by which social networks may affect gendered success. We do so by looking at networks of collaborations to research projects in 10 years of public funding in Italy. We use a permutation t-test analysis to see if men and women, who are equally successful in getting funded for a high number of projects, build equal or unequal collaborative networks, and if these differences vary across disciplines with more or less tenured women. Our methodological contribution proposes an innovative way of measuring brokerage and constraint in one mode projections of bipartite networks, highly clustered and valued, and measures gender homophily by taking into account the higher or lower number of women available for collaborations in different disciplinary sectors. We find that men and women build similar collaborative networks; we also find that the more women are tenured in a discipline, the more they occupy higher ranks and lead research projects. When women's presence and high-rank roles increase, we see that men's collaborative networks are significantly more heterophilous, although women do not seem to significantly increase their collaborations with other women. We then regress the individual measures of network structure and composition against the average amount of money scholars receive per project and we interact each measure with gender, to observe if women obtain similar returns to men by building equal collaborative networks. Our results show that despite the similarities in network configurations in nearly all disciplines women still receive less funding than men. This suggests that an increased representation of women in academic fields is not enough to reduce inequalities. Our results are not directly generalizable to other international contexts, but the mechanisms that favor academic returns and exacerbate inequalities, and our methodological approach, are portable to other scientific and organizational contexts where collaborations are key to produce success.