The study of growth and perceived benefit after traumatic events has been hailed as one of the most promising directions for stress research. This research, however, has been limited by several methodological limitations. These limitations are addressed in this prospective study, which examines perceived benefit and mental health adjustment after 3 different types of disaster. Survivors of a tornado in Madison, Florida, had the highest rates of perceived benefit, followed by survivors of a mass killing in Killeen, Texas, and survivors of a plane crash in Indianapolis, Indiana. Perceived benefit 4-6 weeks postdisaster predicted posttraumatic stress disorder 3 years later. Perceived benefit moderated the effect of severity of disaster exposure on mental health diagnosis change over time. Without perceived benefit, as exposure severity increased, the amount of recovery decreased. If benefit was perceived, as exposure severity increased, the amount of recovery increased. Implications for clinical interventions and future research are discussed.