Eliciting Model Structures for Multivariate Probabilistic Risk Analysis

Notionally objective probabilistic risk models, built around ideas of cause and effect, are used to predict impacts evaluate trade-offs. In this paper, we focus on the use expert judgement fill gaps left by insufficient data understanding. Psychological contextual phenomena such as anchoring, availability bias, confirmation bias overconfidence pervasive have powerful effects individual judgements. Research across a range fields has found that groups access more diverse information ways thinking about problems, routinely outperform credentialled individuals prediction tasks. structured group elicitation, make initial independent judgements, opinions respected, participants consider judgements made others, they may opportunity reconsider revise their estimates. Estimates be aggregated using behavioural, mathematical or combined approaches. contrast, modelers been slower accept host psychological frailties biases afflict parameters events also influence model assumptions structures. Few, if any, quantitative analyses embrace sources uncertainty comprehensively. However, several recent innovations aim anticipate behavioural social in construction mitigate effects. outline approaches eliciting combining alternative effect. We discuss translation into equations assumptions, assessing potential for factors affect models. strengths weaknesses advances structured, group-based accommodate variety understandings