About Crowd Forecasting
Using the “Wisdom-of-the-Crowd” to Estimate Mass Killing Risk
The Early Warning Project uses two types of crowd forecasting to identify countries at risk for an onset of mass killing: an ongoing public opinion pool and an annual comparison survey.
We encourage any interested individuals to participate in these crowd-forecasting projects. Research shows that small crowds can forecast political events more successfully than individuals; expert forecasts are not more accurate than public ones; and larger participant pools produce more accurate forecasts. (See Suroweicki and Tetlock)
Photo above: South Sudan. Jason Patinkin/USHMM.
Public Opinion Pool
An opinion pool is a structured process for collecting and combining opinions (in our case, forecasts) from a group (pool) of people. The larger and more diverse the set of participants, the more accurate we expect the forecasts will be.
Our questions on Good Judgment Open focus on selected countries at relatively high risk for mass killing. Participants can make forecasts throughout the year, and we encourage participants to update their forecasts by reacting to events in the countries in question. Forecasts are aggregated (or “pooled”) on a question-by-question basis. Create an account with GJ Open to participate and see how forecasts have changed over time.
Note: We have paused our public opinion pool while we analyze the data collected to date. Information about any future EWP opinion pools will be posted on this page.
Will an armed group from a country engage in a campaign that systematically kills 1,000 or more of its civilians?
Opinion Pool Data Loading ...
Comparison Surveys
Each December, we conduct a survey on the relative likelihood of countries to experience a new onset of mass killing in the following calendar year. We call this our “Comparison Survey,” since we use a pairwise comparison method: Respondents are presented pairs of countries and asked to choose which is more likely to experience a new mass killing.