Approach 2 Expert Judgements

The weightings in this approach are determined according to experts judgment. The principle of this approach is to submit checklists to a group of experts in road safety. They will be asked to mark the weights for different indicators and to assess the necessity of each indicator subjectively. The results (responses) will be transformed to 

There are different techniques that can be applied in assessing the survey results and opinions of experts such as the Multi-Criteria Analysis (Mendoza et al. 1999) and Delphi Technique (Harold et al. 1975). For now and because of the lack of our survey data, 1 do not discuss these methods further here. 

RSDI range RSDI Rating Less than 20% 10% 20-40% 30% 40-60% 50% 60-80% 70% More than 80% 90% 

This approach depends on the expert panel and it requires a criterion assuming that the expert panels are recognised as fully qualified experts in the area of road safety. Besides, the combined information they have is good enough to judge on the issue we are dealing with. Moreover, it should be clear that it would be difficult to obtain the opinions of experts regularly and frequently. 

---

Table 5.4 Fire Rankings Using Risk Matrix Approach - Expert Judgement...

There is no generally accepted and systematically applied approach for assessing and handling uncertainties in the risk assessment process. It relies heavily on case-by-case expert judgement. This introduces flexibility in the system but also makes it less predictable since the results will to some extent depend on what experts are performing the risk assessment and their particular training, experience and expertise. 

One objective of the GHS is for it to be simple and transparent with a clear distinction between classes and categories in order to allow for self classification as far as possible. For many hazard classes the criteria are semi-quantitative or qualitative and expert judgement is required to interpret the data for classification purposes. Furthermore, for some hazard classes (e.g. eye irritation, explosives or self-reactive substances) a decision tree approach is provided to enhance ease of use. 

The approach to classifying mixtures includes the application of expert judgement in a number of areas in order to ensure existing information can be used for as many mixtures as possible to provide protection for human health and the environment. Expert judgement may also be required in interpreting data for hazard classification of substances, especially where weight of evidence determinations are needed. 

A6.5.12.1 Rating of responses to comprehensibility testing requires expert judgement as to the correctness of the response. Previous experience in Zimbabwe has shown that content analysis of open-ended responses may be feasible where observers are carefully standardized in their approach. 

An index based on the sum or average of bioassay end-points is the simplest to devise. In some instances it may be desirable to combine tests of acute lethality with sublethal tests in order to include a spectrum of organisms and/or responses. Indices are easier to construct if toxicity end-points are first translated into toxic units. The numerical values then can be summed like the chemical properties of a sample. An alternative would be to classify results on an ordinal scale (e.g. 0-10) based on the observed severity of effect. The approach is more subjective, but at least it incorporates expert judgement that should enter the assessment of data at some point. A ranking scale allows any kind of environmental measurement to be included in the index. 

Andrulewicz, E., Kruk-Dowgiallo, L., Osowiecki, A., 2004a. An expert judgement approach to... 

The goal of human error quantification is to produce error probabilities, building on task analysis and error identification techniques to provide a probabilistic risk assessment (PRA). This provides numerical estimates of error likelihood and of the probability of overall likelihood of system breakdown. Quantification of error is the most difficult aspect of HRA, often heavily reliant on expert judgement, rather than the more rigorous approach of actual observation and recording of error frequencies. Such techniques are little used in healthcare but have been successfully applied to anaesthesia (Pate-Cornell and Bea, 1992). Nevertheless, some hospital tasks, such as blood transfusion, are highly structured and the quantification of errors probabilities would seem to be eminently feasible (Lyons et al, 2004). 

How to approach the expert judgements is important since the answers will often depend on how the questions are asked and how the existing information is presented. A number of important details and processes related to applying expert judgements are addressed in Cooke (1991). 

In this case, we regard the initial DU-failure rate as a random variable Apu o with a prior distribution that expresses our uncertainty. Our prior knowledge about the rate of DU-failures may come from OREDA (2002), own previous experience, expert judgement, and so on. A common approach is to assume that our imcertainty can be modeled by a Gamma distribution with parameters a and y (Rausand and Hoyland 2004), with mean and standard deviation ... 

The grey box model approach, i.e., a combination of the black box and white box model is proposed for remaining useful hfe prediction. Use of expert judgement is suggested to compensate for lack of operating ejqrerience and operational data. 

In the safety domain, we also face the same severe limit on the availability of data to quantify failure probabilities. Expert opinion in the form of subjective probabilities has been a dominant source for failure probabilities in quantifying risk in safety science. Cooke (1991) and Cooke and Goossens (2000) presented a structured expert judgment procedure to combine experts opinion in a scientifically defensible way. The overall goal of the method developed by these authors is to achieve rational consensus in the resulting assessments, and thus to enable the information of diverse experts or stakeholders to be incorporated into the process by which the results are reached. We have successfully applied this approach in a series of studies in the field of aviation and chemical industry studies (Ale et al. 2009 2011 2013) with Bayesian Belief Networks (BBNs) quantification. Although the structure expert judgement is not yet implemented in the security world, the scientific proof of such a method can be implemented in this domain. 

The Technique for Human Error Rate Prediction (THERP) was developed by Swain and Guttman (1983) to evaluate the probability of human error within specific tasks. THERP uses a fault tree approach to model Human Error Probabilities (HEP), but also attempts to account for other factors in the environment that may influence these probabilities. These factors are referred to as Performance Shaping Eactors (PSE). The probabilities used in THERP can either be generated by the analyst, usually from simulator data, or can be taken from tables generated by Swain and Guttman from available data and expert judgement. 

Embrey, D.E., Humphreys, R, Rosa, E.A., Kirwan, B., Rea, K. 1984. SLIM-MAUD An approach to assessing human error probabilities using structured expert-judgement. USNRC Report NUREG/CR-3518. Washington, DC USNRC. 

This study also provides an evaluation of the available data and a scientific approach combining feedback data and expert judgement. 

During the abstraction process, critical aspects controlling a given process are identified, conceptual models are developed, parameters are defined, and uncertainties estimated. The PA abstraction is supported by a combination of site and laboratory data, more detailed process models, and in some cases expert judgement (Fig. 10-2). The purpose of this study is to outline approaches that can be used to abstract the results from geochemical sorption models for use in PA. 

The risk matrix approach is a simple subjective method to quantify the probability of occurrence and severity of the associated consequences, however, it lacks a formal way to quantifying expert judgement and opinion when using the risk matrix. This would entail that conflicting opinions of two different analysts on the severity of an accident could result in a deadlock. 

---