The effects of scientific collaboration network structures on impact and innovation: A perspective from project teams

Abstract

Scientific collaboration is critical in tackling complex research challenges, necessitating optimized configurations of research teams. While existing research primarily examines the impact of collaboration network characteristics on the impact and innovation of individual papers, there is less focus on these characteristics within the context of research projects. To bridge this gap, this study adopts the perspective of project teams and explores the influence of scientific collaboration network structures on the impact and innovation of research outputs. By employing ordinary least squares regression and negative binomial regression methods on a dataset encompassing 21,618 NSF grants and their associated 351,550 publications, we rigorously analyze how specific network characteristics impact the innovation and impact of the research outputs. The results reveal a negative correlation between the count of structural holes and both the impact and conventionality of the team's papers. Meanwhile, the small world of a project team positively correlates with the papers' impact and displays an inverted U-shaped relationship with innovation. Further analysis confirms that there is no interactive effect between structural holes and small world. A series of robustness checks have been conducted, demonstrating that these findings are robust. This study contributes valuable insights for scholars, institutions, and policymakers aiming to enhance research team effectiveness. It underscores the nuanced impacts of network properties on research outputs, offering a new perspective by focusing on project-based team structures rather than individual paper collaborations.