northeastern university seal

Volunteers serve a crucial role in post‐disaster situations, providing resources, emotional support, and labor at a time when local and national government capacity may be diminished. The number of volunteers who assist post‐crisis can range from dozens to more than a million. Yet little is known about the broader conditions ‐ geographic, environmental, and otherwise ‐ that drive more (or fewer) volunteers to disaster sites. Using a new dataset of nearly 60 Japanese disasters between 1995 and 2019, we analyze the factors driving volunteer turnout. Controlling for a number of factors, including impact of the disaster, media coverage, and the area touched by the disaster, we find three that correlate most strongly with turnout: the number of dead and missing, the size of the population affected by the shock, and the time period of the year. Moving beyond tables of regression coefficients, we use simulation and graphics to illustrate the relationship between key variables of interest along with the uncertainty about our predictions. These findings ‐ robust across multiple model types, including OLS, generalized linear model, and left censored tobit regressions — bring with them important policy implications for residents, NGOs, and decision makers.


See full paper here.