Research

Analysis

Coding, memos, and interpretation.

AnalysisSelect study
Select a study to begin.

Codebooks & codes

Coding assignments

Analytic memos

Reference & policy
Start analysis →Analysis controls
01Prepared
02Open coding
03Codebook review
04Focused coding
05Theme review
06Interpretation review
07Closed

Analysis modes

Different analytical tasks, one accountable evidence chain.

The workspace supports qualitative and mixed-method inquiry without forcing every study into the same method. Each analytical object remains linked to the study, evidence, analyst, version, and review state that produced it.

Qualitative codingApply controlled codebooks to transcripts, notes, documents, responses, and other working evidence.
Analytic memosRecord interpretations, questions, contradictions, methodological choices, and emerging explanations without treating them as findings.
Comparison matricesCompare participants, groups, sites, time periods, cases, codes, and themes while preserving links back to source evidence.
Theme developmentDevelop and revise candidate themes from coded evidence, memos, negative cases, and competing explanations.
Mixed-method integrationBring structured measures and qualitative evidence together without erasing differences in method, uncertainty, or provenance.
Interpretive reviewTest explanations against the evidence, alternatives, limitations, institutional context, and researcher judgment.

Codebook discipline

Codes should evolve without rewriting history.

A codebook is a controlled analytical instrument. Definitions, boundaries, examples, mergers, splits, retirements, and reviewer decisions remain versioned so later interpretation can be understood in context.

01DefineName the code, analytical purpose, inclusion rule, exclusion rule, and examples.
02TestApply it to evidence, compare difficult cases, and record ambiguity or overlap.
03ReviseChange the codebook deliberately; preserve the prior version and the reason for revision.
04ReviewCheck consistency, contrary cases, analyst assumptions, and whether the code still serves the study question.

Analytical controls

Rigor comes from visible decisions, not hidden certainty.

These controls make it possible to inspect how an interpretation emerged, where it is weak, what evidence challenges it, and which decisions were made by a researcher rather than inferred by a system.

01Source traceabilityCodes, memos, matrices, themes, and interpretations remain linkable to the evidence objects that support or challenge them.
02Versioned codebooksDefinitions, inclusion rules, exclusions, examples, mergers, splits, and retired codes preserve revision history.
03Contrary evidence visibleNegative cases, disconfirming material, ambiguity, and unresolved contradictions are retained rather than silently discarded.
04Evidence and judgment separatedThe workspace distinguishes source material, participant representation, coded observation, analyst interpretation, and proposed finding.
05Pseudonymous by defaultRoutine analysis uses participant identifiers and working copies that exclude direct identifiers unless approved study methods require otherwise.
06AI permission boundaryMachine-assisted analysis is available only where study governance permits it, and human researchers remain responsible for interpretation and release.

Machine-assistance boundary

AI may assist analysis. It does not become the analyst of record.

Where a study permits machine assistance, the system may help organize evidence, suggest candidate codes, compare passages, summarize patterns, surface contradictions, or draft analytical prompts. Those outputs remain provisional. A human researcher decides what counts as evidence, whether a code is valid, how context changes interpretation, and whether any finding is ready to advance.