Views
Description
NUSAP (Numerical, Unit, Spread, Assessment and Pedigree) is designed as a diagnostic and analysis tool of uncertainty in science and policy. It incorporates the quantitative and qualitative dimensions of uncertainty and promotes an extended review process. It is based on the idea to quantify five qualifiers (Numerical, Unit, Spread, Assessment and Pedigree), displaying them in a standardized and self-explanatory way.
-
Numerical is usually expressed as an ordinary number, but can also be presented as a general quantity.
-
Unit can be defined conventionally or with additional information.
-
Spread conveys the spread or uncertainty and thus can be usually computed by one of the methods described above.
These quantitative measures are supplemented by two qualitative quantifiers:
-
Assessment provides a qualitative judgement about the information involved. It could be the systematic error of an analysis or it might be the qualifier “optimistic” or “pessimistic”.
-
Pedigree conveys an evaluation process of information. A set of pedigree criteria is used to assess different aspects.
Subjectivity and arbitrariness has been limited by coding the qualitative judgement for each criterion into a scale linked to linguistic descriptors. An example of such a matrix is given in Risbey et al. (2001). It contains criteria such as “Proxy Representation” to express how good or close a measure of quantity that we measure or model is to the actual quantity we seek or represent; “Empirical Basis”, which refers to the degree which direct observations, measurements and statistics are used to estimate parameters; “Methodological Quality”, which refers to the norms for methodological rigour in this process applied by peers in the relevant disciplines; and validation, which refers to which degree one has been able to crosscheck the data and assumptions used. It has to be noted that some of the criteria of pedigree mentioned are in fact also part of methods to evaluate model uncertainty (e.g. extended Generalized Likelihood Uncertainty Estimation (GLUE) approach). The pedigree scheme has also been extended to incorporate the socio-political context (see references in van der Sluijs et al., 2003).
Software
The website http://www.nusap.net disseminates NUSAP tools including tutorials and papers.
Advantages
Identifies and displays different sorts of uncertainty in quantitative information and allows a clear and transparent assessment of uncertainties. It explicitly acknowledges the issue of quality of information, which is not part of many traditional risk and uncertainty methodologies. Therefore, the open discourse is enhanced and the communication between expert and lay improved.
Disadvantages
Pedigree scores might be misinterpreted. Moreover, motivational bias towards good pedigree scores in elicitation, especially when they concern specific ‘pet’ methods have been reported (van der Sluijs et al., 2003). Many different ways of computing scores could be proposed and a socio-political agreement is required to set standards.
References and Further Reading
The NUSAP website disseminates NUSAP tools including tutorials and papers.
-
Funtowicz, S.O. and Ravetz, J., 1990. Uncertainty and quality in science for policy. Kluwer, Dordrecht. ( Amazon.co.uk)
-
Risbey, J.S., van der Sluijs, J.P. and Ravetz, J., 2001. Protocol for Assessment of Uncertainty and Strength of Emission Data - Nr. E-2001-10, Department of Science Technology and Society, Utrecht University, Utrecht.