Glossary
Sensitivity Testing
The study of how variation in the inputs or assumptions of a model affects its output. Sensitivity analysis is used to assess model robustness, identify influential parameters, and communicate the degree of certainty in results to decision-makers. Techniques range from simple ...
Definition
The study of how variation in the inputs or assumptions of a model affects its output. Sensitivity analysis is used to assess model robustness, identify influential parameters, and communicate the degree of certainty in results to decision-makers. Techniques range from simple one-at-a-time perturbation to global methods such as Sobol indices.
Why It Matters
No model is perfectly specified. Sensitivity testing reveals which assumptions matter most and whether conclusions hold under reasonable alternative assumptions. This is critical in policy analysis, engineering, and clinical research, where decisions depend on model outputs and stakeholders need assurance that results are not driven by arbitrary choices.
Example
A cost-benefit analysis of a transportation project finds a net present value of $12 million under base-case assumptions. Sensitivity testing shows that varying the discount rate from 3% to 7% changes the NPV from $18 million to $4 million, identifying the discount rate as the most influential assumption and flagging that the project may not be viable at higher rates.
Related Terms
Software Notes
- SPSS: No dedicated sensitivity module; use custom syntax or the Simulation module
- R:
sensitivitypackage for Sobol indices;morris()fromsensitivityfor Morris screening - Stata:
senspecfor sensitivity/specificity analysis; custom scripts for general sensitivity