The process of reaching agreement can surface neglected categories

What it is

When coders sit down to argue through disagreed items, the point is not only to lift a coefficient. Negotiation as category discovery treats “why did we land on two different codes” as an analytic event. It often exposes a missing class of phenomena, a dimension that had been collapsed, or two different referents in the material that each coder had grabbed. Agreement is a by-product; the named gap is what this step is for.

Why it happens

Independent coding forces each person to cut the material once, without social pressure. The disagreements that follow are therefore not noise but two incomplete readings colliding on the same stretch. One common collision: one coder caught the action, the other the emotion or the explanation, and the manual had only one cell. Another: the stretch literally satisfies two code definitions, which means the categories are not exclusive, or a higher theme is needed. A third: both feel “neither fits,” which often points to an unnamed residual class. If negotiation ends at “this sounds smoother,” that information is sanded off.

Studying it

Record the negotiation: verbatim stretch, two initial codes, each person’s reason, the final code, and whether a definition was added, split, or rewritten. When analyzing those records, count “new code,” “double reading found,” and “unitizing error” separately; do not keep only the final codebook. Audit whether final agreement is bought by one side yielding; systematic yielding means power, not material, is deciding categories. New codes born in negotiation must be tested on material that was not in the meeting, or they only fit the items just discussed.

Where it stops holding

Not every disagreement deserves a new code. Some are missed lines, fatigue, or unfamiliarity with project terms; the fix is more training, not an inflated codebook. Negotiation also cannot replace independent coding: talking from the start produces spuriously high agreement and kills the discovery function, because two readings were never allowed to form. Majority vote under time pressure shortens discovery into a poll.

Related

• Same group: Q4.11.1 Agreement needs a shared codebook and explicit definitions · Q4.11.2 Low agreement coefficients diagnose ambiguity in the coding frame · Q4.11.4 High agreement is not analytic value
• Adjacent: Q4.01 Affinity diagramming · Q4.02 Coding and thematic analysis
• Search terms: consensus coding · negotiated agreement · category discovery