Uncertainty in policy relevant sciences

Martin Paul Krayer von Krauss

Research output: Book/ReportPh.D. thesisResearch

110 Downloads (Pure)


In addition to possessing the expert knowledge and technical know-how required to provide public services, engineers and scientists function as an important source of legitimization for regulatory decisions. According to the liberal philosophy upon which the regulatory process is based, regulatory authorities should only intervene in instances where development could lead to harmful effects. Their decision to intervene should be based on facts, ideally considered within the framework of a rigorous rational methodology (e.g. risk assessment) so as to ensure that their interpretation of the facts is as objective as possible. This can be quite problematic in situations where scientific knowledge is limited, facts are uncertain, the stakes are high and values are conflicting.

In many cases, the complexity of the regulatory issues scientists and engineers are asked to study far surpasses that of typical laboratory problems. Knowledge is often limited and the facts are uncertain. That which has commonly been referred to under the umbrella term ‘uncertainty’ actually hides important technical distinctions. The well-established notion of uncertainty as a statistical or probabilistic concept leaves out many important aspects of the uncertainty encountered when assessing complex policy problems, such as the uncertainty generated by assumptions and ignorance of causeeffect relationships.

The precautionary paradigmhas emerged in the context of the above realisations. Under this paradigm, the role of scientists and engineers is to participate in the collective effort of producing, evaluating and applying knowledge, considering the interests at stake, and making a necessarily provisional decision. Formal methods are required for experts to assess uncertainty, and these methods must transcend the notion of uncertainty as a statistical or probabilistic concept. Ultimately, they should help foster a reflexive and humble attitude towards development.

This dissertation illustrates how a novel conceptual framework for uncertainty analysis, the Walker & Harremoës (W&H) framework, can be applied. The W&H integrated uncertainty analysis framework synthesizes a variety of scholarly contributions on uncertainty, in order to provide an interdisciplinary theoretical framework for systematic uncertainty analysis. Uncertainty is broadly defined as being any deviation from the unachievable ideal of completely deterministic knowledge of the relevant system. The framework distinguishes between three fundamental dimensions of uncertainty: the location, level and nature of uncertainty.

The W&H framework was applied to analyse two case studies related to the risk assessment of genetically modified crops: i) the risk of developing “super weeds” through the use of herbicide tolerant rapeseed; and ii) the phenomena of transgene silencing. As the experts involved were not familiar with the W&H framework, expert elicitations were used to communicate the W&H framework to the experts in such a way that their knowledge of uncertainty was obtained, without them being overly intimidated or confused by the novelty of the concepts they were presented with.

The results obtained indicate that, not withstanding efforts to clarify the relationships between the concepts put forth in the W&H framework, experts did not use these concepts consistently. Nonetheless, the approach was successful in making explicit levels of uncertainty deeper than the statistical uncertainty commonly reported. As is the case for the concept of “risk”, different people will have different (subjective) perspectives on the concept of “uncertainty”. The approach successfully revealed that there are a variety of perspectives on the uncertainty that characterizes the cases studied. Thus, although the results yielded by studies such as the ones presented here may seem ambiguous, they are a valuable contribution to the discussion of the quality of the information underlying regulatory decisions.
Original languageEnglish
Place of PublicationKgs. Lyngby
PublisherDTU Environment
Number of pages76
ISBN (Print)87-89220-97-8
Publication statusPublished - Feb 2006


Dive into the research topics of 'Uncertainty in policy relevant sciences'. Together they form a unique fingerprint.

Cite this