UnCoDe 'UNcertainty and COnfidence in safety arguments: effect on expert DEcision makers'

Staff and Funding

Principal Investigator: Prof Lorenzo Strigini

Contact People: Prof Lorenzo Strigini

Funding for City, University of London: £239,175

Funding Source: Leverhulme Foundation

Duration: October 2010 - March 2013

Project Description:

This interdisciplinary project seeks progress in understanding and improving complex decision making, involving communication of uncertainty among experts, as it takes place in regulatory approval for complex systems with high societal impact (high benefits, high potential risk).

We address the two, often conflicting, requirements that arguments for approval should be rigorous and complete, describing uncertainty in rigorous mathematical terms, but they should lend themselves to be processed correctly by the human mind. Our focus is on the needs of the decision makers as consumers of these complex arguments.

The chosen application area is probabilistic safety claims concerning software, an area that is difficult and contentious. For example, it plays a crucial part in the uncertainties about approval of the next generation of nuclear reactors.

We aim to produce both new knowledge about the cognitive processes involved in this kind of decision making, and concrete aid for producing better "safety cases".

Multidisciplinary work in this project involves:

• probabilistic reasoning for reliability and safety, especially for software, and contributions to the practice of safety cases in industry.
• empirical and theoretical research on risk perception and communication, judgmental forecasting, decision making under risk and uncertainty.

City, University of London Personnel: Prof Lorenzo Strigini (Principal Investigator), Prof Robin Bloomfield, Prof Bev Littlewood, Dr Eugenio Alberdi, Dr David Wright, Prof Peter Ayton (Psychology Department)