Subjective Probability

Subjective probability is a measure of the degree of belief held for the truth of an answer to a question. It is used in the quantification of uncertainty due to lack of knowledge, also called epistemic uncertainty. The word epistemic stems from the Greek word for knowledge. It indicates that this uncertainty has its origin in the nature and limits of knowledge. In the case of epistemic uncertainty, there are several answers to a question that are considered as possibly true, while there is only one true answer. The true answer will either be deterministic or probabilistic, depending on the formulation of the question. Both may be subject to epistemic uncertainty. A probabilistic answer uses probability in its classical frequentistic interpretation to quantify uncertainty due to random (or stochastic) variability, also called aleatory uncertainty.

Decision making involves asking questions. The answers to most of the questions will be subject to epistemic uncertainty. Quantitatively expressing this uncertainty by subjective probability enables one to employ all concepts, methods, and tools from probability calculus in the quantification of the combined influence of the uncertainties on decision making.

The following paragraphs explain by example the difference between subjective probability and probability in its classical frequentistic interpretation. Then, the connection with medical decision making is briefly pointed out. Rules for calculation are followed by a short introduction into the specification of subjective probability values. The discussion ends with two practical examples and a summary of how subjective probability is used in the uncertainty and sensitivity analysis of results from decision models.