The impact of network embeddedness on research output

This paper aims to further our understanding of how embeddedness affects the research output and impact of scientists. The analysis uses an extensive panel data that allows an analysis of within person variation over time. It explores the simultaneous effects of different dimensions of network embeddedness over time at individual level. These include the establishment of direct ties, the strengths of these ties, as well as the density, structural holes, centrality, and cross-disciplinary links. Results suggest that the network dynamics behind the generation of quality output contrasts dramatically with that of quantity. We find that the relational dimension of scientists matters for quality, but not for output, while cognitive dimensions have the opposite effect, helping output, while being indifferent toward impact. The structural dimension of the network is the only area where there is some degree of convergence between output quantity and quality; here, we find a prevalence for the role of brokerage over cohesion. It concludes by discussing implications for both network research and science policy.