 |
| HomeAbout SRAMembersEventsNewsletterJournalsOpportunitiesResourcesContact |
| Back to 2009 Meeting Main Page |
Society For Risk Analysis Annual Meeting 2009
Risk Analysis: The Evolution of a Science
Session Schedule & Abstracts
* Disclaimer: All presentations represent the views of the authors, and not the organizations that support their research. Please apply the standard disclaimer that any opinions, findings, and conclusions or recommendations in abstracts, posters, and presentations at the meeting are those of the authors and do not necessarily reflect the views of any other organization or agency. Meeting attendees and authors should be aware that this disclaimer is intended to apply to all abstracts contained in this document. Authors who wish to emphasize this disclaimer should do so in their presentation or poster. In an effort to make the abstracts as concise as possible and easy for meeting participants to read, the abstracts have been formatted such that they exclude references to papers, affiliations, and/or funding sources. Authors who wish to provide attendees with this information should do so in their presentation or poster.
Common abbreviations
W3-F |
| Chair(s): Mark Burgman |
|
W3-F.1 13:30 Towards a Universal Deca-Scale for Risk Assessment. Krishnamurthy N*; (Self-Employed) safety@profkrishna.com Abstract: Risk is commonly assessed as a qualitative estimate derived by combining likelihood of occurrence of a hazard with severity of its consequences, each from three or more levels, via a ‘risk matrix’. Defining likelihood and severity in words leads to vagueness in level assignment, and wide variations between different industries and projects. Comparison of procedures and statistics across industries, and across countries with different languages is largely infeasible. Absolute quantification of likelihood and severity is practical only in a few cases where every input and output is fully quantifiable. The best compromise is ranking of levels by giving numbers 1, 2, … to increasing likelihood and severity. Risk, taken as the product of the level ranks, is also a number. Field professionals are more comfortable with numerical assessments. Standardisation of these ranks is very important for consistency of data and robustness of conclusions. Only a common worldwide ranking scheme would enable universal application and evaluation – which is not possible now. Author proposes a ‘Deca-Scale’ (1 to 10) for components of workplace risk, similar to Richter Scale, covering credible ranges of likelihood and severity in human experience. He explofres different measures of risk likelihood and severity in Deca-Scale. As risk index now goes from 1 to 100, it automatically gives percentage measure of risk in a global context. Author presents three practical examples of Deca-Scale applications. It would need collaboration of experts from various specialities to evaluate and agree upon common Deca-Scales for workplace activity. As humans are well accustomed to fixing a value for everything on a ‘scale of 1 to 10’, it should be possible to initiate efforts to develop Deca-Scales in a concerted worldwide effort. Who knows, apart from getting all of us to talk the same language, in due course this may also raise safety culture, improve risk assessment and reduce accidents around the world! |
|
W3-F.2 13:50 Experience with Encoding Expert Opinions. Paté-Cornell E*; Stanford University mep@stanford.edu Abstract: In this talk, I will describe good and less good experiences with encoding and using expert opinions, based on three studies that I have done in the past: patient risk in anesthesia, risk of an accident in a shuttle mission caused by a failure of the tiles, and seismic risk analysis. 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 the set of implicit assumptions on which the expert is sometimes relying without telling the analysts. I will also show that whereas one can obtain pretty good data about the center of a distribution, it is much more difficult to generate that information for the tails, unless the problem has been carefully designed to decompose extreme scenarios in a way that lends itself to the analysis of rare, dependent events, the combination of which is sometimes called "perfect storms". |
|
W3-F.3 14:10 A new approach to model cost and duration of projects in stochastic environments. Hajbagheri M*, Yaraghi N; Royal Institute of Technology (KTH) mansour.hajbagheri@gmail.com Abstract: This study is build upon the earlier works around management of construction projects and introduces several adjustments to the existing methods. These adjustments offer some advantages over the existing methods. Firstly, existing methods are either based on historical data or expert ideas. Since historical data lacks the ability to project the future events and expert ideas are subjective in nature, using one of them is not adequate. Therefore, the proposed adjustment is to use both subjective and objective source of data in order to benefit from both sources and increase the accuracy and reliability of input to the Monte Carlo simulation. Secondly, existing models use deterministic values for weightings. To avoid the subjectivity associated with deterministic weightings, fuzzy logic is deployed to derive the required values when needed. Thirdly, this study utilizes Delphi technique for reducing variation among the expert ideas and obtaining the most accurate result out of it. To verify the advantage of the proposed method over the existing ones, it is applied to the empirical data obtained from a real construction project. Results show a significant improvement in the accuracy and reliability of cost and time estimates. |
|
W3-F.4 14:30 Making better use of experts. Burgman M*, Speirs-Bridge A, Fidler F, McBride M; University of Melbourne markab@unimelb.edu.au Abstract: It is well established that expert judgements are often inaccurate, overconfident and biased by a host of uncontrolled psychological and contextual influences.Yet, there is no broadly established procedure for engaging with experts to estimate facts in risk analysis. This presentation summarises the results of recent experiments on the a priori assessment of expert status, the reliability of expert judgements, and methods for combining expert judgements in ecological risk analysis. The results suggest a structured protocol that, if adopted routinely, could substantially improve the quality of expert judgements. |
[back to schedule]