One Epidemic, Many Estimates (1EME)

"Many analyst" projects give teams of researchers the same data and the same research questions, then see how they use the methods of their choice to arrive at different answers to those questions. Prior to the workshop, there will be a “many analyst” collaboration (see a previous example here). Teams will be asked to analyse the mortality of the 1918–20 influenza pandemic in the US populations of Kentucky and rural Maryland. They will be asked to estimate between 3 and 6 quantities. These quantities will relate to 1) the absolute level of pandemic mortality across space, 2) the relative level of pandemic mortality between social groups (defined by gender, race, and geography), and 3) the timing of the pandemic.  

Cleaned, highly granular death count and denominator data will be provided to teams. Death count data will be available by date of death, county of death, (imputed) cause of death, and single year of age for both states. In Kentucky, death counts will also be available by the deceased’s registered gender, and in Maryland, they will be available by the registered race. 20 years of death count data will be available from Kentucky, and 10 years of data will be available from Maryland. For those who want to incorporate it into their model, granular weather data and shapefiles will also be available.

Teams can comprise up to 5 researchers. Participants will be asked to submit their estimates, reproducible code that generates those estimates (in any programming language), and a written description of their estimation technique (between 250 and 1,000 words).

Data will be available to teams starting on 15 September 2025 and all submissions are due by 15 March 2026. You can already sign up to participate in the collaboration using the link below!

The workshop will begin with a systematic discussion of the agreement and disagreement between teams’ estimates and the effects of the explicit and implicit analytical choices they made. This will be followed by roughly 8 sessions on key topics in excess mortality methodology, spread across the afternoon of the first day and morning of the second day. Most of these sessions will comprise half-hour lectures followed by open discussion. The afternoon of the second day will turn to building consensus regarding the best practices for estimating particular estimands. 

The current plan for the discussion sessions is as follows:

1. What are we estimating when we estimate excess mortality?
2. How do we estimate our estimands?
3. What data do we need for accurate and precise estimation?
4. How do we quantify uncertainty in excess mortality estimates?
5. What data challenges do epidemics cause?
6. Frontiers in methodology: Spatial modelling
7. Frontiers in methodology: Advanced forecasting techniques
8. Frontiers in methodology: Out-of-sample prediction

Participants are welcome to suggest additional ideas for sessions, or to volunteer to give (all or part of) the lecture on a topic. 

Teams can register to participate in the many analyst collaboration here:  

Registration to attend the workshop (by analyst teams and other researchers) will begin in early 2026. 

• Hampton Gaddy (LSE)
• Monica Alexander (University of Toronto)
• Jennifer Dowd (University of Oxford)
• Eric Schneider (LSE)
• Maarten van Wijhe (Roskilde University)

 Contact Hampton (h.g.gaddy@lse.ac.uk) if you have any questions.