• In a three-year prospective study in which life events and psychiatric symptoms were assessed every two months, Fourier analyses led to the classification of event-symptom relationships into five typologies. Only one event-symptom typology, representing 9% of subjects, was consistent with the commonly held causal model. The results indicate that future progress in life-events research will require prospective designs, more sophisticated methods for gathering life-events data than the Schedule of Recent Experiences, focus on qualitative features of events (eg, undesirability or threat), and identification of personal characteristics and coping styles that might augment or attenuate the health impacts of life happenings.