Research activities in the Risk perception area.
Quantitative and Qualitative Assessments of Risk
A major problem in academic analysis of risk is the disconnection between quantitative and qualitative methods. Traditionally, science and engineering emphasises quantitative methods and regards these are superior to qualitative methods. This sometimes has the unfortunate consequence of promoting unqualified quantification – what has been termed GIGO science (garbage-in, garbage-out). In response Jerome Ravetz and Silvio Funtowicz developed the NUSAP process for verifying and checking on the quality of numerical data used in risk assessments and computational modelling for policy purposes.
Economics and cognitive psychology are traditionally quantitative approaches, hence have approached risk perception and responses using quantitative methods, including statistical and simulation modelling. The interpretative social sciences have been much less quantitative, relying more upon in-depth qualitative methods such as interviews, focus groups, in-depth discussion groups, ethnography and so on.
Some efforts are now underway to bridge the gap between ‘technical quantitative’ and ‘social qualitative’ research styles. While a single ‘best approach’ to the problem has yet to emerge, some promising lines of enquiry are:-
- Agent Based Modelling (ABM): rule-based approach to understanding decision-making at agent-scale; emergent properties emerge through interaction between agents, giving rise to (frequently unexpected) behaviours and responses.
- Other simulation modelling approaches such as System Dynamics.
- Multi-criteria mapping – whereby decision makers select weights and scores across a range of criteria which they also choose for deciding between different options for responding to a given policy problem.
- Policy laboratories – whereby key decision makers come together to explore their responses to different scenarios / eventualities, these decisions being quantified in some way during the process, which can then be used in a decision-making modelling.