ABSTRACT
The present paper focuses on the use of expert knowledge to construct HRA models. Rigorous treatment of expert judgment is seen as necessary before tackling the integration of the three main sources for HRA (data, theoretical models, judgment). The first aspect addressed by the present paper is the aggregation of expert estimates to inform the quantitative relationships of a BBN (the Conditional Probability Distributions, CPDs). In particular, the Bayesian approach recently presented in Podofillini & Dang (2013) is adopted, which allows formally and transparently aggregating expert estimates, so that both the inherent variability of the human error probability and the experts’ variability are represented and distinguished in the final probability distribution.
