Decisions about coronavirus response are necessarily based on statistical models of prevalence, transmission risks, case fatality rate, projection of future spread of infection, and estimated effects of medical and social interventions. Much of this modeling and inference is being done using the Bayesian framework, an approach to statistics that is well suited to integration of information from different sources and accounting for uncertainty in predictions that can be input into decision analysis. This is a project to develop computing tools to make Bayesian methods more accessible to researchers in quantitative social science and public health who are studying COVID-19 and epidemic models more generally. This work promises to advance scientific knowledge by enabling researchers to fit more flexible and realistic models accounting for multiple sources of uncertainty in data, and to advance societal goals by facilitating more accurate and granular estimates of exposure, reproduction rate, and other aspects of epidemic spread that inform public and private decision making. This project also provides professional development opportunities for a post-doctoral researcher, as well as student training.
The research will be done in the open-source programming language Stan, which has already been used in several influential COVID-19 models as well as in economics, political science, biology, political science, and many other application areas. Specifically, the project includes: documentation and language features to make Stan programs easier to write and evaluate; continuation and extensions of existing collaborations on mathematical models for epidemic spread, causal models for estimating policy effects, and survey adjustment; and improved implementations for differential-equation models, which serve as the core of models for disease transmission and other diffusive social and biological processes.
This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.
