Crossing boundaries: institutional collaboration and its impact on scientific performance
摘要
Interinstitutional collaboration is critical for scientific performance. In current innovation system, the interactions among different types of institutions are increasingly strengthening, promoting scientific development through resource complementarity and knowledge exchange. This study argues that understanding the institutional type composition in cross-sector collaboration is of great significance for scientific performance. Utilizing a large-scale dataset, we undertake a comprehensive investigation of eight institutional types involved in academic collaboration, exploring the relationship between the institutional composition and both citation impact and disruption. The results reveal that the seven non-academic institutions play distinct roles in enhancing citation performance and disruptiveness of research, and most of them are generally beneficial for scientific impact and innovation. We further analyze the function of institutional type diversity in scientific collaboration, and find that as the diversity of institutional types increases, the citations of papers show an upward trend, however, the relationship between institutional type diversity and the disruption of papers exhibits an inverted U-shape pattern. Interestingly, we find that the scale of institutional collaboration effectively moderates the nonlinear relationship between institutional type diversity and scientific impact and disruption. We also observe field heterogeneity in the role of institutional participation and the relationship between institutional collaboration and scientific performance. The study further reveals the heterogeneous effects of different institutional types under varying levels of institutional type diversity. When institutional type diversity is high, certain types of institutions consistently have a positive impact on citation impact and disruption in multi-type collaboration, while others struggle to utilize their potential in enhancing scientific performance. Our findings provide valuable insights for policy-making, research management, and practical team formation.