Fair Comparison of Alternatives
A practical framework for comparing feasible alternatives fairly, consistently, and transparently.
Build a Fair Comparison
Start with a fair basis
A comparison is only useful if each option is judged against the same decision and objectives. First state the problem and the outcomes that matter. Then include a clearly described : what would happen under business as usual, or if only the minimum were done. Without that reference point, it is hard to tell what an option changes.
Compare only credible, feasible alternatives—not merely variations of a favored proposal. Before ranking them, choose criteria that follow from the objectives and help distinguish the options. Apply the same time period, population, assumptions, and cost perspective to every option. If one option is assessed over a different period or for a different population, apparent differences may reflect the comparison setup rather than the options themselves.
Organize the evidence, not just the scores
Use a shared set of criteria to make trade-offs visible. The table is a template: for each option, record what is expected, what it requires, what could hinder delivery, and who benefits or bears costs. In the evidence column, note sources, limitations, assumptions, and uncertainty. This keeps costs and resources, implementation risks, and distribution alongside expected outcomes instead of hiding them in one overall score.
| Criterion | Option A | Option B | Evidence and confidence |
|---|---|---|---|
| Expected outcome | What is expected? | What is expected? | Source, limitations, uncertainty |
| Cost and resources | What is required? | What is required? | Estimates, perspective, time period |
| Risks and implementation | What might hinder delivery? | What might hinder delivery? | Evidence and assumptions |
| Distribution | Who benefits or bears costs? | Who benefits or bears costs? | Affected groups and locations |
Keep observed evidence distinct from . Report how strong the evidence is, whether it is relevant to the local context, and what remains uncertain. A precise-looking estimate is not necessarily a certain one: when evidence is weak, say so rather than presenting the number as settled.
When effects cannot be measured reliably in numbers, or values and lived experience matter, a structured qualitative comparison can still help. Use consistent labels such as low, medium, and high, and give a brief reason tied to evidence. Define the labels consistently across options. Record disagreements and uncertainty instead of forcing artificial consensus.