This seminar is organized by LIEPPs Discriminations and category-based policies research group.

January 28th. 12:30-13:45.

Sciences Po, room CS16, 1 place Saint Thomas d'Aquin, 75007 Paris. 

Speaker

Wei Cheng, visiting professor at LIEPP

Abstract

This paper examines how the diffusion of important new scientific ideas depends jointly on the career age of innovators, the characteristics of potential adopters, and the collaboration networks that connect them. Using comprehensive data on U.S. biomedical research from MEDLINE, we identify major new ideas introduced between 1980 and 2008 through large-scale natural language processing of article titles and abstracts. We then track how these ideas are adopted over the subsequent decade by other researchers working in related fields, measuring diffusion through coauthorship networks.

The analysis highlights the importance of adopting a “two-sided” perspective on idea diffusion that accounts for both innovators and potential adopters as well as their network proximity. We show that the relationship between innovator age and idea adoption varies sharply with network distance. At short distances—among close collaborators—ideas introduced by young researchers are adopted most intensively. At larger network distances, however, ideas introduced by mid-career researchers are adopted at higher rates. Because most potential adopters are located at greater network distances, overall adoption of new ideas is highest for innovations introduced by researchers in early to mid-career stages.

These patterns arise primarily from differences in network position over the career life cycle. More senior and mid-career innovators are more centrally embedded in collaboration networks, which enables their ideas to reach a larger and more distant set of potential adopters and supports broader diffusion. This network advantage is especially important for mid-career researchers, whose ideas tend to spread most widely. In contrast, early-career innovators are typically embedded in smaller networks, but this disadvantage is partly offset by their proximity to younger potential adopters, who are more willing to adopt new ideas.

The paper further shows that young researchers play a disproportionate role in identifying and adopting valuable ideas that might otherwise be overlooked, especially when those ideas are introduced by nearby young innovators or published in less prominent journals. These findings underscore the importance of early-career scientists in the diffusion process, not only as innovators but also as adopters who help promising ideas gain traction.

Finally, the authors conduct simulations to assess how the ongoing aging of the scientific workforce affects idea diffusion. They find that aging reduces the overall adoption of new ideas through two channels of comparable magnitude: a decline in the quality of ideas produced by older innovators and a decline in receptivity among older potential adopters. Focusing on only one side of the diffusion process therefore understates the full impact of workforce aging.

Overall, the paper provides new evidence on how age, networks, and receptivity jointly shape the diffusion of scientific ideas, and highlights the critical role of network structure and career dynamics in determining whether important discoveries achieve their full impact.