Model uncertainty, the COVID-19 pandemic, and the science-policy interface

Abstract

The COVID-19 pandemic illustrated many of the challenges with using science to guide planning and policymaking. One such challenge has to do with how to manage, represent and communicate uncertainties in epidemiological models. This is considerably complicated, we argue, by the fact that the models themselves are often instrumental in structuring the involved uncertainties. In this paper we explore how models ‘domesticate’ uncertainties and what this implies for science-for-policy. We analyse three examples of uncertainty domestication in models of COVID-19 and argue that we need to pay more attention to how uncertainties are domesticated in models used for policy support, and the many ways in which uncertainties are domesticated within particular models can fail to fit with the needs and demands of policymakers and planners.