Alarmist decisions with divergent risk information

Receipt of multiple sources of risk information ideally should foster sounder decisions under uncertainty. This paper's original survey results for environmental risks suggest that the learning process is reasonable in many respects, bur it does not accord with a rational Bayesian learning model. Divergent risk assessments from different sources produce extreme violations of rationality, as there is inordinate weight on the high risk assessment. This alarmist reaction holds for both government and industry information sources. This phenomenon may account for the commonly observed phenomenon of public overreaction to highly publicised risks.