The generation of probabilities

Summary In principle, probabilities are – at least in practical decision situations – of a subjective nature. They cannot be detected objectively but have to be elicited from people. People are not used to making precise probability statements and are frequently reluctant to do so. However, in order to support the making of important decisions under uncertainty, probabilities should be quantified numerically. Probabilities are elicited by direct or indirect queries. An essential tool is the decomposition of complex events into basic events for which the probabilities are easier to determine. With the aid of the Bayes-Theorem, prior probabilities and additional data are transformed into posterior probabilities. This is a very useful tool as humans are not intuitively capable of carrying out correct transformations. A number of inadequacies of intuitive human information processing distort the process of probability generation. On the one hand, these deficiencies concern the usage of incomplete or improper data. On the other hand, incorrect probability judgments result from the wrong intuitive processing of probabilities. A third source of error is insufficient willingness to critically view one’s own probability judgments.