Abstract: Conventional statistical and regression analyses of safety incidents ignore the prospect for underreporting. In settings where underreporting is anticipated, the use of conventional models may lead to incorrect conclusions. ...