SYNOPTIC ABSTRACT
Subjective probability was developed for representing uncertainty in real world situations. Traditionally, it is viewed either as an intuitive notion that can be measured directly, or as a by-product of an agent's preferences among uncertain choices. This paper argues that subjective probability is often not a primitive notion, but a construct that must be built using the available information. In this view, probability represents the extent to which available information supports the occurrence of an event. Approaches to probability construction are discussed, and one of the models is illustrated with examples. The discussion also sheds some light on the sources of difficulties in the measurement of subjective probability.
Key Words and Phrases: