Methods & Frameworks

18 articles

Qualitative coding is a family of methods, not a single technique, and choosing among them is the first methodological decision of any language-analysis project. These articles cover the major approaches—thematic analysis, grounded theory, content analysis, framework analysis—along with the deductive-versus-inductive choice that shapes everything downstream, from codebook design to what counts as a finding.

Alongside the general frameworks sit the specialized coding systems that clinical research runs on: motivational-interviewing process codes like the MISC and MITI, communication taxonomies like VR-CoDES, and the protocols that turn a therapy session or medical consultation into utterance-level data. The mechanics matter here—unitizing, memo writing, hierarchical code structures, and the codebook development cycle that keeps a scheme usable after its authors move on.

Throughout, the emphasis is practical: what the method assumes, where it breaks, and how to document the process well enough that a reviewer—or a second coder—can follow what you did.

Two rising speech-shapes on a horizontal sequence line—one faint and dotted, one solid—each meeting an aperture that either widens or narrows, illustrating VR-CoDES cues versus concerns and providing versus reducing space.

Methods & Frameworks

VR-CoDES Coding System: Cues, Concerns, and Responses

How the VR-CoDES coding system codes patient emotional cues versus concerns and the provider response, using the providing-space vs reducing-space axis.

Jul 22, 2026Read

Two parallel segmented speech-waveform ribbons, one per language, aligned span for span except a single coral bracket sitting at a different height, illustrating a severity anchor whose meaning shifts across languages.

Methods & Frameworks

Localize a Rating Scale: Annotating in a Second Language

Localize a rating scale for coding in a second language: what back-translation validates, what it misses, and how to re-anchor the coding examples.

Jul 22, 2026Read

A single labeled tile multiplies across a grid, each column stacking into taller bars, while a small branching path splits into parallel coder lanes that converge at a marked total on the right.

Methods & Frameworks

Annotation Cost Calculator: Budget and Timeline Math

An annotation cost calculator: turn items, labels per item, time, wage, redundancy, and QA overhead into a labeling budget and timeline. See the math.

Jul 21, 2026Read

Two diverging colored coding paths over a segmented speech waveform converge onto a single labeled span once a small dated decision-log card is placed between them.

Methods & Frameworks

How to Write Annotation Guidelines That Reduce Disagreement

Write annotation guidelines that reduce annotator disagreement: operational definitions, positive and negative examples, and an edge-case decision log.

Jul 21, 2026Read

A single speech waveform read three ways—a plain self-dial, a finely-notched clinician's caliper, and a bracketed tagged span—illustrating self-report, clinician-rated, and transcript-based measurement of the same interview.

Methods & Frameworks

Clinician-Rated vs Self-Report Scales: A Methods Trade-off

Clinician-rated vs self-report scales compared: who rates, the bias each carries, reliability and cost trade-offs, and transcript coding as a third path.

Jul 21, 2026Read

One large diagnostic-label shape unfolding into many small bracketed, color-tagged spans over an abstract transcript band, with a single coral-highlighted tag—showing a diagnosis resolved into observable annotatable phenomena.

Methods & Frameworks

DSM Criteria Annotation: From Diagnosis to Taggable Spans

Turn DSM-5 and ICD-11 criteria into observable, taggable phenomena: a worked MDD-to-PHQ-9 mapping and why you annotate symptoms, not a diagnosis.

Jul 21, 2026Read

A segmented speech waveform tipping a two-pan balance—more change-talk tiles on one side, fewer sustain-talk tiles on the other—illustrating that the balance of client language, not the raw count, carries the predictive signal.

Methods & Frameworks

Change Talk vs Sustain Talk: Coding Client Language

Change talk vs sustain talk, coded: how the MISC classifies client language, why the balance predicts outcomes, and how to segment it reliably.

Jul 21, 2026Read

Two facing reading lenses examine the same speech-transcript ribbon: one resolves it into the client's bracketed change-and-sustain-talk tags, the other into the clinician's counted behavior tallies, with a single coral spark at the utterance both capture.

Methods & Frameworks

MISC vs MITI: Coding the Client vs the Clinician in MI

MISC vs MITI, disambiguated: the MITI codes the clinician's MI fidelity; the MISC codes the client's change and sustain talk. Compare purpose and use.

Jul 21, 2026Read

A balance scale weighing two stacks of tally-marks against each other beside two proportioned bars in a roughly two-to-one ratio, illustrating MI-consistent versus MI-inconsistent behavior counts and the reflection-to-question ratio.

Methods & Frameworks

MITI Motivational Interviewing: A Fidelity Coding Guide

How MITI motivational interviewing coding works: behavior counts, four global scores, and the R:Q ratio and %complex reflections that grade MI fidelity.

Jul 21, 2026Read

A small tray of a few labeled sample tiles feeds a loop that measures, forks into two coder paths, converges, and only then opens a gate to a much larger grid of unlabeled tiles.

Methods & Frameworks

How to Run a Pilot Annotation Round Before You Scale

Run a pilot annotation round before you scale: scope, sample size, double-coding, measuring agreement, and revising the codebook. See the checklist.

Jul 21, 2026Read

A vertical rating scale with tick-marks expands rightward into a grid of codebook entry cards, each showing a color tag, a definition line, and an include/exclude mark, one card highlighted in coral.

Methods & Frameworks

Rating Scale to Codebook: Operationalize Any Instrument

A step-by-step method to go from rating scale to codebook: turn each item into a code, anchors, examples, and decision rules coders apply consistently.

Jul 21, 2026Read

A patient's speech bubble opening beside a six-rung response ladder, illustrating the ECCS opportunity-and-response structure.

Methods & Frameworks

The Empathic Communication Coding System, Explained

The Empathic Communication Coding System codes empathy in clinical talk as a two-step sequence: patient opportunities, then a 0–6 provider response ladder.

Jul 15, 2026Read

A story arc drawn as a ribbon in six segments with the evaluation high-point marked, illustrating Labov's narrative structure.

Methods & Frameworks

Labov and Waletzky Narrative Analysis: The Six Parts

Labov and Waletzky narrative analysis breaks an oral story into six parts. See each part, why evaluation is the heart, and how reliably it can be coded.

Jul 15, 2026Read

Five funnels from wide-open to narrow, illustrating the NICHD prompt continuum from invitation to suggestive.

Methods & Frameworks

The NICHD Protocol: Coding Interviewer Prompt Types

The NICHD protocol classifies a forensic interviewer's prompts from open to suggestive. See the five types, why open prompts win, and how to code them.

Jul 15, 2026Read

A patient-involvement gauge reading low, illustrating that observed shared decision-making is consistently low.

Methods & Frameworks

The OPTION Scale: Measuring Shared Decision-Making

The OPTION scale rates how far a clinician involves the patient in a decision. See the behaviors it scores, why real scores are low, and how to code it.

Jul 15, 2026Read

Speech-bubble shapes sorting into five distinct channels, illustrating Searle's five classes of speech act.

Methods & Frameworks

Searle's Speech Acts: The Five Illocutionary Types

Searle's speech acts sort every utterance into five illocutionary types. See the classes, direction of fit, examples, and the tie to dialogue-act tagging.

Jul 15, 2026Read

An argument as a supported scaffold with one dashed ghosted beam for the implicit warrant, illustrating the Toulmin model.

Methods & Frameworks

The Toulmin Model of Argumentation: The Six Elements

The Toulmin model of argumentation breaks an argument into six functional parts. See the elements, a worked example, and why warrants are hardest to code.

Jul 15, 2026Read

A speech waveform segmented into colored, bracketed spans over a faint punch-card texture, illustrating clause-level content analysis that tags and scores spoken language.

Methods & Frameworks

Gottschalk-Gleser Content Analysis: Coding Affect in Speech

How Gottschalk-Gleser content analysis scores affect in speech, clause by clause—the seven depression subscales, the weighted coding, and its NLP legacy.

Jul 13, 2026Read