Skip to content

Sensitivity Analysis Protocol

Vary key assumptions or parameters to see which ones materially change the conclusion.

The Diagnostic Story

Symptom: A recommendation looks precise and settled, but the assumptions feeding it are rough, contested, or undocumented. Stakeholders argue about the conclusion without examining which assumptions would actually change it. Small cosmetic variations leave the result unchanged, yet no plausible stress range has ever been tested. A single best-estimate answer is being presented as if it were more certain than the evidence warrants.

Pivot: Make uncertain assumptions explicit, vary them across defensible ranges or scenarios, measure how the conclusion responds, and surface the critical assumptions and flip points — the inputs where a plausible change would change the recommendation — so validation effort and monitoring can be concentrated there.

Resolution: Decision confidence becomes conditional and auditable: stakeholders know which assumptions the conclusion is fragile to, validation resources are directed at those inputs first, and the recommendation is accompanied by the stress range rather than presented as a single number. Fragile conclusions become visible before commitment rather than after.

Reach for this when you hear…

[financial modeling] “The IRR looks great at the base case, but what happens if volume is 20% lower — does the whole project flip negative?”

[climate policy] “The cost-benefit ratio changes by a factor of three depending on the discount rate, so we can't just present one number.”

[clinical decision support] “That threshold was set when the false-positive rate was estimated at 5% — what's the recommendation if it's actually 15%?”

When This Archetype Applies

Partial catalog groundingSome structural conditions are represented by existing abstractions, but no sufficient condition set is fully represented.

A model, analysis, plan, or decision recommendation looks settled, but its conclusion may depend heavily on uncertain assumptions whose influence has not been made visible.

What this problem means

The structural problem is conditional confidence. A plan, model, forecast, ranking, or recommendation appears settled, but its conclusion may depend on uncertain inputs whose influence is hidden. This creates two bad options: either actors overtrust the baseline answer, or they reject the analysis wholesale because uncertainty is present.

Sensitivity analysis creates a middle path. It keeps the baseline explicit but treats it as conditional. The question becomes: under what plausible assumption changes does the conclusion remain stable, and under what changes does it fail?

Show the applicability expression

Applicability expression2 distinct conditions

Assumption-sensitive decisionandUnknown influential changes
Algebraic12

groundedpartly groundedopen

2 conditions, all required.

2Required in every casenumbered 1–2

These hold no matter which pattern applies.

1

Assumption-sensitive decision · grounded

A decision depends on forecasts, estimates, weights, thresholds, or uncertain causal assumptions.

2

Unknown influential changes · open

Small changes in context could plausibly change the outcome, but no one knows which changes matter.

Other requirements and context (3)

Why these sit outside the expression

Supporting contextit may accompany or help interpret the situation, but it is not a load-bearing condition in a sufficient diagnostic set.

  • Supporting contextStakeholders disagree about which assumptions are plausible or important.

  • Supporting contextThe model output is being used to justify a costly, risky, irreversible, or politically contested commitment.

  • Supporting contextA single best-estimate answer is being presented as more certain than the evidence warrants.

1 of 2 conditions grounded · 1 open.

Read the methodologyDownload the trigger-logic data

Mechanisms / Implementations

  • One-way Sensitivity Analysis: Moves one input at a time across its range while holding everything else fixed, then ranks assumptions by how far each alone swings the outcome.
  • Two-way or Multi-way Sensitivity Analysis: Varies two or more inputs at once across a grid of combinations to expose interaction — the effects that appear only when assumptions move together.
  • Scenario Variation: Bundles many assumptions into a few internally coherent named worlds and reads the outcome under each to judge whether the plan survives all of them.
  • Threshold Analysis: Solves backward for the exact value of an input at which the recommendation flips, turning that break-even point into a monitoring trigger.
  • Tornado Chart: Draws each input's outcome swing as a horizontal bar, sorted widest-first, so the dominant drivers are legible at a single glance.
  • Sensitivity Table: Records one row per assumption — its range, outcome response, materiality verdict, and critical flag — so the whole analysis can be audited line by line.
  • Probabilistic Sensitivity Simulation: Draws thousands of joint samples from input distributions and reports the share of draws in which the recommendation holds versus flips.
  • Assumption Stress-test Workshop: Convenes the people who own or dispute the assumptions to argue defensible ranges, name the decision-carrying ones, and set the validation agenda.

Abstractions this archetype builds on — directly (a source ingredient) or as a related pattern. Links follow the typed catalog namespace.

Built directly on (3)

Also references 9 related abstractions

Variants

Narrower or domain-specific specializations that share this archetype's core structure. Recognized variants are established; candidate variants are provisional.

One-way Sensitivity Protocol · parameter isolation variant · recognized

Test conclusion fragility by varying one uncertain parameter at a time while holding the rest of the baseline fixed.

Threshold Sensitivity Analysis · decision flip variant · recognized

Identify the parameter value or condition at which a decision, ranking, constraint, or recommendation changes.

Scenario Sensitivity Protocol · scenario bundle variant · recognized

Test a conclusion against coherent bundles of assumptions rather than varying inputs independently.

Probabilistic Sensitivity Protocol · uncertainty distribution variant · recognized

Represent uncertain inputs probabilistically and estimate how often the conclusion changes under sampled combinations.

Editorial Notes

Problem Classification

Classification: Uncertainty, Evidence & Inference FailureForecast, Scenario, Assumption & Sensitivity Uncertainty

Problem kernel: conclusions hide sensitivity to uncertain assumptions

Rationale: Earliest causal condition: A model, analysis, plan, or decision recommendation looks settled, but its conclusion may depend heavily on uncertain assumptions whose influence has not been made visible.

Independent corroboration: The earliest necessary condition in the frozen evidence is: A model, analysis, plan, or decision recommendation looks settled, but its conclusion may depend heavily on uncertain assumptions whose influence has not been made visible. That is a forecast scenario assumption and sensitivity uncertainty problem because Plans and forecasts hide assumption dependence, alternative futures, reference-class evidence, revision conditions, and uncertainty ranges.

Review outcome: Independent reviewer agreement; high confidence.