Ecologist Charles Hall argued that "the real strength of the models is in their sensitivity analysis (where one could examine the response of the model to parameters or structures that were not known with precision, and in the examination of the behavior of the model components relative to that of the real system in question (i.e., validation). By undertaking sensitivity analysis and validation, a great deal can be learned about the real system, including what you do not know." (Hall, 2020) We revisit Hall's conjecture with some example, looking at how sensitivity analysis has established itself as a discipline with many practitioners in several disciplines. We look at dos and don'ts, at some best practices, and at some simple ones. We conclude with some mention of what to do where a 'technical' sensitivity analysis is not enough.
Join at: imt.lu/aula2