Engineering risk analysis and the use of expert opinions

In this talk, I will describe first the methods that we use in engineering risk analysis, based on systems analysis and (generally Bayesian) probability.  These methods originally developed in the nuclear power industry have been applied to many other fields, especially for cases in which we do not have sufficient statistics and need to decompose the problem to use all available information.  I will present some of the applications that have been developed in my research team, including space systems (the tiles of the space shuttle) and medical procedures (anesthesia).

Expert opinions are a key part of the data that we use.  I will present and discuss good and less good experiences with encoding and using expert opinions in several of these studies.  I will show how the quality of these data depends in good part on the way the problem has been formulated, the precision of the question, the familiarity of the expert with the domains (and the size of the sample on which he/she is drawing), but also on the set of unstated assumptions on which the expert sometimes implicitly relies.  The skills of both the expert and the analyst are particularly critical in the analysis of extreme scenarios that include rare, dependent events, the combination of which is sometimes called “perfect storms”.

Presenter(s)

Elisabeth Paté-Cornell is the Burt and Deedee McMurtry Professor in the School of Engineering and has been chair of the Department of Management Science and Engineering at Stanford University since its creation in January 2000.  She is a member of National Academy of Engineering, of the boards of Trustees of Aerospace Corporation and of InQtel, of the NASA/JPL Technical Advisory Committee, and of the board of advisors of the Naval Postgraduate School.  She was recently a member of the President’s Intelligence Advisory Board (2001-2008). She is a past president and a fellow of the Society for Risk Analysis and a fellow of the Institute for Operations Research and Management Science.  Her primary areas of teaching and research are engineering risk analysis, risk management, and decision analysis.  Her research has focused on the inclusion of both technical and organizational factors in probabilistic risk analysis models, with application to the management of the tiles of the space shuttle, offshore platforms, and anesthesia during surgery.  In her recent work, she has used game and risk analyses to address some national security and counter-terrorism problems as well as management issues in industry.