Model Challenges - The Covid-19 Model Challenges

The main goal of this project is to study the collective capacity of social scientists to predict future cross-national and sub-national patterns of COVID-19 mortalities using political and socio-economic variables. We provide data on pre-existing political and social variables that may be causally determinative of COVID-19 deaths for countries across the world and also for subnational units within the United States, India, and Mexico. The first round of the Model Challenges asked social scientists to submit statistical models to predict future cumulative mortality for these four sets of units. Submitters are invited to provide “general” models (with parameters to be estimated on data on 31 August 2021) and “specific” models with specific parameter guesses provided by submitters. Contributors of the 10 models that receive the most weights from a stacking exercise in each challenge will be asked if they wish to be included as co-authors. A second round of Model Challenges will open in January 2022 and will target undergraduate and graduate students enrolled in statistics and data analysis courses. It will ask entrants to provide models to predict future cumulative mortality as of August 2022.

This project benefits from a multi-institutional collaboration and networking. We are particularly proud of the fact that the Steering Committee comprises all levels of seniority, from graduate students through chaired professors. They are:

• Alex Scacco (Project co-chair), Social Science Center Berlin (WZB)
• Macartan Humphreys, WZB and Columbia University
• Kim Dione, University of California at Riverside
• Alberto Diaz-Cayeros, Stanford University
• Eva Vivalt, University of Toronto
• Sampada KC, University of British Colombia
• Eugenia Nazrullaeva, London School of Economics and Political Science