Consistencies and Biases in Risk Perception: I. Anchoring Process and Response-Range Effect

49 undergraduate subjects provided intuitive frequency judgments of deaths due to 11 well-known causes. The responses were collected in three different formats, estimates of the frequency of deaths per 10,000 people in the population, estimates of the frequency of deaths per 100 people, and rank-order of the causes according to the estimated frequencies. Analysis showed that estimated proportions were reliably higher in the “out of 100” format than in the “out of 10,000” format. I argue that processes of anchoring and adjustment underlie this phenomenon. Also, the subjects' response tendencies seemed to maintain consistency across different methods of judgment. Relationship with previous risk-perception research is discussed.