What is it about?
Sensitivity analysis is critical to gauge the relevance and plausibility of mathematical or statistical models. Yet sensitivity analysis is either overlooked or performed unsatisfactorily. In practice most sensitivity analyses are made as to 'tickle' the potential uncertainty of the model prediction, instead of probing it thoroughly. Hence it would not off the mark to say that most published sensitivity analysis are false.
Featured Image
Why is it important?
A modelling work without a sensitivity analysis is like an orthopedic diagnosis without an X-ray.
Perspectives
We see sign of improvements in the take up of global sensitivity analysis, very unequal across different disciplined - the worse is economics in case the reader wonders! Journals could play a role to improve responsible use of quantitative information.
Professor Andrea Saltelli
University Pompeo Fabra, Barcelona School of Management
Read the Original
This page is a summary of: Trends in sensitivity analysis practice in the last decade, The Science of The Total Environment, March 2016, Elsevier,
DOI: 10.1016/j.scitotenv.2016.02.133.
You can read the full text:
Resources
Sensitivity analysis: Scope and limitations
Lesson given to European Commission staff at the course "Statistical and participatory tools for Impact Assessment" 28-29 April, 2016, Brussels (BE).
Sensitivity analysis: An introduction
Presentation given at the Universitat Autònoma de Barcelona UAB, Bellaterra, July 18 2016.
This article in plain English
Why most published sensitivity analysis are wrong. A one page illustration of what sensitivity and uncertainty analysis mean and why most SA seen in the literature are wrong.
Contributors
The following have contributed to this page
