Whose uncertainty assessments (probability distributions) does a risk assessment report: the analysts' or the experts'?

It is often said that the aim of risk assessment is to faithfully represent and report the knowledge of some defined experts in the field studied. The analysts' job is to elicit this knowledge, synthesise it and report the results as integrated uncertainty assessments, for example, expressed through a set of probability distributions. Analysts' judgements beyond these tasks should not be incorporated in the uncertainty assessments (distributions). The purpose of the present paper is to discuss the rationale of this perspective. To conduct a risk assessment in practice the analysts need to make a number of judgements related to, for example, the choice of methods and models that to a large extent influence the results. And often the analysts are the real experts on many of the issues addressed in the assessments, in particular, when it comes to understanding how various phenomena and processes interact. Would it then not be more appropriate to fully acknowledge the role of the analysts as uncertainty assessors and probability assigners, and see the results of the risk assessments as their judgements based on input from the experts? The discussion is illustrated by two examples.

► Expert elicitation does not yield only expert knowledge.
► Expert elicitation incorporates assessor knowledge.
► Use of elicitation results must consider information content.