Coercion through air power has been and continues to be an important tool o
f foreign policy. Multivariate probit analysis is used to test three hypoth
eses on all instances of air power coercion from 1917 to 1999: (1) air powe
r coercion attempts are more likely to work if they exploit military rather
than civilian vulnerabilities, (2) the regime type of the target affects t
he chances of success, and (3) success is less likely if the attacker deman
ds that the target change its leadership. Results show that coercion is mor
e likely to work if the target's military vulnerability is higher, but high
er levels of civilian vulnerability have no effect on the chances of coerci
on success; that target regime type has no effect; and that success is less
likely when the attacker demands the target change its leadership.