A well-substantiated, surprising finding is that people judge the occu
rrence of an event of low probability as less likely when its probabil
ity is represented by a ratio of smaller (e.g., 1 in 20) than of large
r (e.g., 10 in 200) numbers. The results of three experiments demonstr
ated that the phenomenon is broadly general and occurs as readily in p
re- as in postoutcome judgments. These results support an interpretati
on in terms of subjective probability, as suggested by the principles
of cognitive-experiential self theory, but not an interpretation in te
rms of imagining counter-factual alternatives, as proposed by norm the
ory.