C1 ยท Model: structuring
Turn the categories into a structure
C1 asks whether the coding scheme still holds after use across the collection. It repairs the weaknesses found so far, then groups the third-order themes into fourth-order themes.
Where this stage starts
B4 supplies settled placements and their disagreements. The aggregator also has the pilot results, the categories readers often combined, and the themes left unplaced. Together these show where the scheme worked and where its boundaries hid the material.
What the aggregator does
The aggregator revisits troublesome categories and a seeded sample of placements. It can split a category that combines different actions, merge categories with no useful distinction between them, add a missing category, or retire one that does no work. Each change needs a reason and a record of the material it affects. A split requires rereading, because the old label cannot say which new category an item belongs to.
Multiple placements need care. An act may perform two functions, or two categories may be poorly separated, and reader confidence cannot settle which. The scheme is not changed to make readers agree with its author.
The aggregator also returns to the first-order codes and the participants' language. It looks at how turns are formed and sequenced, how ways of speaking vary, and what participants seem to take for granted. This pass can reveal patterns that earlier labels hid.
Finally, C1 groups the third-order themes into fourth-order themes, each of which states what several categories have in common. Each group needs an argument and boundaries, and no number of groups is set in advance. The stage also marks which ideas come from existing work, which were adapted, and which were named here.
What the next stage receives
C2 receives a revised coding structure and a complete, reconciled placement record. C1 supplies the parts and their grouping; C2 asks how the parts relate in an explanation of organizing.
What the output files contain
Markdown files hold prose; JSONL files hold one structured record per line.
The revised rulesscheme_v2.md, scheme_v2.jsonl
scheme_v2.md explains the revised scheme;
scheme_v2.jsonl records it for tools. These are complete
versions, so later stages read the rules from one file.
Every placement, with the changes kept separateassignments_final.jsonl, assignments_final.count.txt, assign_revisions_only.jsonl
assignments_final.jsonl contains the full placement
record. assignments_final.count.txt records its reconciled
count. assign_revisions_only.jsonl lists only the changes,
so a reviewer can see what C1 altered, with the whole record kept in the
first file.
The fourth-order themesfourth/fourth.jsonl, fourth.md
fourth/fourth.jsonl records the fourth-order groups and
their members. fourth.md explains why each group's members
belong together and what the group means.
What a direct reading of language revealedform_pass-C1.md
form_pass-C1.md records observations about turn
structure, language and assumptions that the earlier category labels
might hide.
Reasons and the next steptier_history.jsonl, notes-C1.md, handback-C1.md
tier_history.jsonl records changes across levels of the
scheme. notes-C1.md explains choices and unresolved issues.
handback-C1.md tells the next stage what changed, what was
checked and what still needs judgment.
Full stage description
Read the complete technical card for exact tasks, reading boundaries, checks and closing rules. The explanation above is a guide; the card holds the instructions.
Show how the stage is divided into its 1 legs
C1a: pass. Repeats: yes. Interview point: end.
Source versions and hashes identify the exact technical documents reproduced here.