Social preferences during major disasters

For centuries, scholars have asserted that prosocial behavior should increase when others face greater harm. No prior study, however, has causally tested how the actual presence of high-level harm influences prosocial behavior, because randomly inflicting high-level harm is unethical. We overcome this challenge by using natural experiments to examine social preferences during major disasters. In the largest causal investigation of prosocial behavior to date, we analyze how plausibly random major disasters (hurricanes, natural fires, mass shootings) causally influence 1,976,649 prosocial behaviors by 14,383 volunteer crisis counselors. Surprisingly, in contrast to theoretical predictions and laypeople’s intuitions, we find that major disasters cause prosocial behavior to fall. This overall effect is driven by inexperienced volunteers; among highly experienced volunteers, major disasters instead cause prosocial behavior to rise.