Perceptions of death rates from hazards described in terms of their occurrence rates (time) versus the number of persons affected (population) were examined in three experiments. In experiment 1, participants judged the death rates from five anonymous real hazards that were described in terms of time or population, using cognitive, affective, and behavioral scales. They rated the time-related hazards to be significantly more frequent and more fearful. In experiment 2, the hazards were given specific names. Participants rated them on the same scales as those used in experiment 1 and described their impressions in response to 9 pairs of adjectives. Again, hazards described by time were rated more frequent and more fearful. The mean scores of the three factors determined through factor analysis of qualitative responses also differed significantly between time and population. In experiment 3, time- and population-related hazards were judged differently, although numerical values of death rates were controlled for to eliminate the anchoring effect. The psychological reasons for this finding and its practical implications for risk communication were discussed.