Policy Scenario Sensitivity Analysis — Robustness Testing of Policy Models Across Defined Futures

Standard sensitivity analysis varies inputs within a single model configuration. PSSA explicitly structures the analysis around multiple policy scenarios, so sensitivity profiles are computed per scenario and then compared across scenarios to find what is robust and what is scenario-conditional.

Not necessarily. PSSA can be applied to any quantitative model — analytical, spreadsheet-based, or simulation — as long as inputs can be varied systematically and outputs computed. Simulation models are common because they often represent complex dynamics, but the method is not limited to them.

The choice depends on model complexity and computational budget. Morris screening is efficient for model ranking with many inputs. Sobol variance-based indices provide rigorous quantitative decomposition but require more model runs. One-at-a-time analysis is simple but misses interaction effects.

There is no fixed rule, but policy analyses typically use two to six scenarios to keep the comparison interpretable. Scenarios should represent genuinely distinct policy alternatives or futures, not minor parameter shifts, which belong in the sensitivity analysis layer.

Yes, through exploratory modeling approaches (related to robust decision-making). In that case, input ranges are defined by plausibility bounds rather than probability distributions, and sensitivity analysis identifies which regions of the input space change the policy ranking.