Öberg, T.
Lecture at the Annual Meeting of the Society for Risk Analysis, Boston, Massachusetts, December 7-10, 2008.
Abstract
Variability and uncertainty are key concepts in all quantitative risk
assessments. Here, we will describe and discuss how these concepts can be
introduced with practical examples and problem solving, in university teaching
of environmental risk analysis. Our discussion will particularly focus on the
methods found in two popular calculation and simulation software – Crystal
Ball® and Risk Calc™ – that are used in our master’s program at the
University of Kalmar, Sweden. Multimedia models for exposure assessment are easy
to implement with both software, although the separation of variability and
uncertainty can be realized with fundamentally different approaches. Likewise,
dependency assumptions and incomplete information is often treated differently.
We have found it very useful for students in the advanced classes to experiment
with these different approaches to model the same or similar problems. These
exercises bring further insight into statistics, probabilities, modeling, and
the many assumptions involved in probabilistic exposure models. Additionally,
the different approaches to treat uncertainty – intervals, probability
distributions, and ‘p-boxes’ – raise interesting questions regarding the
boundaries of knowledge.
|
|
Swedish homepage