Collaborate but not compete? Gender, homophily, and the productivity-citation gap in inventor networks

As scientific collaboration becomes increasingly global and complex, understanding the network mechanisms underlying gender disparities in innovation is essential. While prior research has often examined single disciplines or national contexts, this study integrates social network analysis and regression modeling to explore how gender, homophily, and network position interact to shape inventive outcomes. Using a 30-year panel of 3,871,017 U.S. patents (1990–2019) and 2,631,166 inventors, we distinguish between star inventors – those in the top productivity and citation percentiles – and non-star inventors to assess whether network advantages operate uniformly across performance strata. Results reveal three key patterns. First, women occupy less central positions and experience productivity gaps in aggregate models. Second, even with star status, top-performing women receive fewer citations than men, indicating that exceptional output does not offset systemic disadvantage. When network–gender interactions are introduced, the female–star coefficient loses significance, suggesting that the observed productivity premium among female stars stems primarily from positional advantages rather than star status itself. Third, men’s networks exhibit stronger homophily, whereas women rely on heterophilic ties – beneficial for productivity but less so for visibility. These findings extend theories of cumulative advantage, showing that structural capital alone cannot neutralize inequities in innovation systems.