Interval Analysis Versus Probabilistic Analysis

B. Garrick

DOI: https://doi.org/10.1111/j.1539-6924.2010.01360.x

2010-03-01

Risk Analysis

Abstract:The Aven paper suggests restricting probability analysis to “objective probabilities.” This implies that there is a clear distinction between so-called subjective and objective probabilities. It also implies the need for a level of precision that may only be useful from an academic and research perspective. The preferred approach is to define probability in such a way as to be reasonably accountable for all of the supporting evidence, if not the subtleties and fine structure of variability and future events. In particular, let the supporting evidence determine the shape of the distribution (an important piece of information) and avoid arbitrary assumptions such as uniform distributions or other measures such as intervals and means, which do not transmit as much information as the full curve. A definition that works well for organizing evidence to support decision making is to consider probability as being synonymous with “credibility.” Thus, probability is a number between 0 and 1 that expresses the degree of credibility of the hypothesis in question, based on the totality of relevant evidence available.(1) As to “subjective” versus “objective” the late physicist E. T. Jaynes(2) said it best: “A probability assignment is ‘subjective’ in the sense that it describes a state of knowledge rather than any property of the ‘real’ world, but is ‘objective’ in the sense that it is independent of the personality of the user. Two rational beings faced with the same total background of knowledge must assign the same probabilities.” The earlier probability definition and Jaynes’s interpretation thereof, together with Bayes theorem, the fundamental principle governing the process of inferential reasoning, provide the foundation for meaningful probabilities for scientific and engineer-