The Structure of Research Teams and Scientific Outputs

Date: 2021-10-22 19:23:40
Time: 20:30-22:00
Speaker: XU Fengli
Category: Talk & Lecture

Speaker: XU Fengli

XU Fengli is a postdoc fellow in the Knowledge Lab and Mansueto Institute for Urban Innovation in University of Chicago. His research interest lies in computational social science and science of science. He currently focuses on understanding and modeling the collective intelligence of scientific teams.

Venue: ZOOM


Meeting ID:871 2559 2985



The advance of science today is characterized by a productivity dilemma—the fast growth in the number of scientists and publications but slow expansion of new concepts and ideas. This diminishing returns of science has triggered concerns on whether science is stagnant. Here we reveal a related social phenomenon—the increasing dominance of “tall teams” working on evaluating existing ideas, as “flat teams” who are more capable of searching for new ideas, diminish in prevalence. We analyze nearly 96,000 author contribution statements from papers published in four reputable journals over 15 years, including Nature, Science, PNAS, and PLOS ONE. We identify two team roles, “brain” and “muscle,” who contribute to idea search—designing research and writing papers, and evaluation—implementing experiments, analyzing data, drawing figures, etc., respectively. As designing and writing activities produce references, we proposed and verified “knowledge exchange index,” a novel measure that identifies “brains” who actively contribute to the collective references based on their individual knowledge. In this way, we infer and analyze the effect of flat and tall teams in 18 million scientific papers published in the past 70 years.