CDB Seminar - Prof Carl Bergstrom, University of Washington

Abstract: Scientific researchers may be driven by curiosity, but they are constrained by the realities of the scientific ecosystems in which they operate and motivated by the incentives with which they are confronted. We can use mathematical models of the research enterprise to understand how scientific norms and institutions shape the questions we ask, the efficiency with which we work, and the discoveries we make about the world around us. In this talk I present two models, both of which were developed in collaboration with statistician Kevin Gross at NCSU. In the first vignette, we consider the efficiency consequences of grant proposal competitions, in which research funding is allocated by a contest mechanism with grant proposals as the outputs. We illustrate the huge inefficiencies of this approach and consider alternative mechanisms including partial lotteries for funding. In the second, we look at how peer review filters — ex ante review as for grant proposals and ex post review as for completed manuscripts — shape the types of questions that researchers pursue. We show that ex ante and ex post review select for different kinds of work, we discuss strategies that scientists may take to navigate both types of review, and we look at how ex ante review discourages risky research projects. Our scientific norms and institutions are not god-given; we create and maintain them. If we can understand their consequences, we have the potential to nudge the norms and institutions in directions better tailored to our contemporary research questions and technologies.

Hosts: Tom Dufor and Matthew Bostock

Zoom:
https://ucl.zoom.us/j/98811478589?pwd=ZXBOR0QyV0dycGtYenp5bUMxUEt6dz09

Meeting ID: 988 1147 8589
Passcode: 901767