How did households interpret chernobyl?. A bayesian analysis of risk perceptions

This paper reports the first attempt to apply a Bayesian learning model to explain how households' perceptions of the risks of nuclear power changed after the Chernobyl accident. The analysis includes information on a sample of suburban Boston households' attitudes toward nuclear power plants before and after the accident. It also uses information on their current and retrospective risk perceptions. The results, based on the stated risk perceptions, appear to suggest that households have difficulty distinguishing the probability of these types of infrequent, and potentially large, events from the severity of the outcomes involved.