AO-19 Ranking uncertainty analysis (stochastic multicriteria acceptability analysis; SMAA; rank acceptability index)
Samples weights and score uncertainty many times and reports how often each option takes each rank, giving rank acceptability indices that show how stable a multi-criteria ranking is.
What it does
Samples criteria weights and score uncertainty many times and counts how often each option takes each rank. Artefact: rank acceptability indices and central weights per option. Read them as how far the ranking depends on the assumed weights and scores, not as the chance that an option is best.
When to use it
Use when options have been scored against weighted criteria, or the material supplies scores and weights, and the decision depends on how stable the ranking is.
When it does not fit
Have the options already been scored on weighted criteria, and does the decision depend on whether the ranking would survive different weights? If yes, use this after the scoring.
Result format
The result is a table with one row per option, giving the rank acceptability index for each rank and the central weights. It states how far the ranking depends on the assumed weights and scores, and does not describe the index as the chance that an option is best.
Requirements on the material
| Property | Requirement type | Requirement |
|---|---|---|
| Content type | Always applies | requires=tabular |
| Unit of analysis | Always applies | requires=record |
| Level of detail available | Always applies | requires=aggregate tables |
| Appraises source credibility | Applies in some cases | no |
Not restricted on: Structure, Variable types, Time structure, How the sample was chosen, Number of sources, Whether sources can be joined, Amount of material, Known data problems, Producer interest, Source dependence, Origin, Document genre, Language, Personal information, Security classification, Licence.
| Minimum material | At least two options, scores on at least two criteria, and a range or distribution for each score or weight. |
|---|---|
| Additional sources needed | none |
| Not for (rules) | none |
| Not for (text) | Rank shares are read as the chance that an option is truly best; they show only how the ranking responds to the assumed uncertainty. |
Choosing and ordering
| Kind of technique | Primary technique, runs in step 5.3 |
|---|---|
| Phase tag | analyse |
| Other phase tags | validate |
| Kind of finding | option ranking; risk |
| Question answered | which option |
| Strongest statement it supports | describe |
| Needs as input |
|
| Makes as output |
|
| Must exist first | None |
| Loop target | return to analyse |
Library checks by stakes
Each stakes level includes the checks of the level below it. Checks that need a person apply at high stakes only, so the process leaves them out at lower stakes (decision 38). Repeatability: repeatable with fixed settings.
| Stakes | Check | Person involved | Counts as a different method |
|---|---|---|---|
| exploratory | assumption and limitation statement | none | no |
| exploratory | code execution for arithmetic | none | no |
| decision-informing | assumption and limitation statement | none | no |
| decision-informing | code execution for arithmetic | none | no |
| decision-informing | repeat runs | none | no |
| decision-informing | sensitivity analysis | none | no |
| high-stakes | assumption and limitation statement | none | no |
| high-stakes | code execution for arithmetic | none | no |
| high-stakes | repeat runs | none | no |
| high-stakes | sensitivity analysis | none | no |
| high-stakes | second technique with different assumptions | none | conditional |
| high-stakes | independent review | needed | no |
Failure modes and limits
| Model suitability | model plus code |
|---|---|
| Typical model failure | Samples weights evenly across the whole range, so an extreme weight on one criterion is treated as likely as a balanced set, and then reports the share of samples as a probability. |
| Error sources | input data; statistical assumptions; expert opinion |
| Certainty limits | Certainty drops when scores come from one rater (risk of bias) and when the weight ranges were set without agreement (imprecision). |
Related techniques
- pairs with AO-12 Multi-criteria decision analysis (MCDA): multi-criteria scoring and ranking uncertainty read the same scores and weights, and the second shows how stable the first ranking is
- pairs with TC1-17 Sensitivity analysis (robustness check, leave-one-out): sensitivity analysis varies one input at a time while this technique varies all weights together
Evidence and tag basis
Documented. SMAA is a published family of methods for multi-criteria decision problems with uncertain or missing information, with rank acceptability indices, central weights and confidence factors as outputs. (Lahdelma and Salminen, SMAA chapter abstract; CRAN smaa package manual)
Tag basis: input profile inference; selection inference; sequencing inference; robustness inference. All tag values are inference unless a source is named.