Team Production, Scientific Competition and Interdisciplinary Research

Abstract

We conceptualize a research project as a CES production function whose factors are sub-team efforts, each one in a distinct field of science. We abstract from the intra-team coordination and free-riding issues to focus on the team decision problem when it competes in a Tullock contest to win a given prize, or say a publication slot in a top journal. We show that the degree of interdisciplinarity of the team outcome at equilibrium is tied in a certain way to the coefficients of the production function, in essence the nature of the problem addressed. Under some mild conditions, interdisciplinarity goes in the reverse direction than the probability to participate in the contest, the production outcome and the probability to win to contest. These results are consistent with scientific research being mostly mono-disciplinary, and scientific impact decreasing with interdisciplinarity. It can also explain that observed interdisciplinary research actually may have larger impact because of the selection bias, not as an outcome of interdisciplinarity per se. It also suggests that outcome is less sensitive to interdisciplinarity when the team is of higher quality. We then bring this model to the data, testing its predictions on an original dataset of nearly eight hundred thousand authorship participations of nearly thirty thousand French professors and researchers over period (1999-2014).