DM-06 Affinity diagram (KJ method)
Groups many recorded items, such as survey comments or interview quotes, into clusters that form from the relationships among the items, and produces labelled clusters.
What it does
Groups many recorded items, such as survey comments or interview quotes, into clusters that form from relationships among the items. Artefact: labelled clusters; read cluster size, number of clusters and orphan items.
When to use it
Use when you have many comments or quotes and want themes to emerge from the items without fixed categories.
When it does not fit
One catch-all cluster dominates, or items move between clusters when a second person sorts them.
Result format
The result is a set of labelled clusters, each holding the recorded items (such as survey comments or interview quotes) that belong together. It gives the number of clusters, the size of each cluster, and the orphan items that joined no cluster.
Requirements on the material
| Property | Requirement type | Requirement |
|---|---|---|
| Content type | Always applies | requires=text |
| Unit of analysis | Always applies | requires=document; requires=claim |
| Appraises source credibility | Applies in some cases | no |
Not restricted on: Structure, Variable types, Time structure, How the sample was chosen, Level of detail available, 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 20 items recorded individually. |
|---|---|
| Additional sources needed | none |
| Not for (rules) | none |
| Not for (text) | One catch-all cluster dominates, or items move between clusters when a second person sorts them. |
Choosing and ordering
| Kind of technique | Primary technique, runs in step 5.3 |
|---|---|
| Phase tag | analyse |
| Other phase tags | verify and prepare |
| Kind of finding | pattern; gap |
| Question answered | what |
| 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: judgement dependent.
| Stakes | Check | Person involved | Counts as a different method |
|---|---|---|---|
| exploratory | spot check | optional | no |
| decision-informing | spot check | optional | no |
| decision-informing | human review sample | needed; left out below high stakes | no |
| decision-informing | prompt variation | none | no |
| high-stakes | spot check | optional | no |
| high-stakes | human review sample | needed | no |
| high-stakes | prompt variation | none | no |
| high-stakes | independent review | needed | no |
| high-stakes | second technique with different assumptions | none | conditional |
Failure modes and limits
| Model suitability | model alone |
|---|---|
| Typical model failure | Forces items into clusters named after the first items seen and leaves orphan items unreported. |
| Error sources | input data; coder or rater judgement; language model |
| Certainty limits | No inter-rater reliability evidence; cluster membership is not set in advance, so reliability depends on a second sort. Risk of bias from the sorter. |
Related techniques
- substitutes for MM-14 Document clustering (computer-assisted clustering; k-means on document vectors): Hand clustering replaces document clustering when volume_class is readable by a person
- pairs with MM-15 Computer-assisted theme building (reading-assisted thematic analysis; human-guided topic labelling): A person reads and names computer-generated groups to build themes
Evidence and tag basis
Documented. Group membership is not set in advance; no inter-rater reliability evidence found. (Wikipedia, Affinity diagram)
Tag basis: input profile inference; selection inference; sequencing inference; robustness inference. All tag values are inference unless a source is named.