Impacts of disaster on optimistic biases about personal risk and on a wide range of other cognitions and affects were investigated. Residents of communities hit by severe tornadoes who suffered no damage or injury (N = 724) and residents of control communities (N = 287) were interviewed shortly after the events. Optimistic biases were strong in control towns, but they were substantially lower, though not eliminated, in the impact towns. Several hypotheses that might explain the decrease in risk biases were examined but were not supported. Variables indicating preoccupation with the threat (frequency of thoughts, vigilance, intrusive thoughts) stood out as the only variables closely tied to individuals' personal experiences during and after the tornado. Judgments of tornado risk magnitude may be constructed on the spot to answer survey questions; whereas preoccupation measures may better reveal respondents' actual ongoing disaster experience.