document-analysis

name	document-analysis
description	Use when analyzing documents, artifacts, policies, or media as qualitative data sources.

Document and Artifact Analysis

Documents include policies, emails, meeting minutes, curricula, clinical records (where permitted), social media threads, photographs, videos, and organizational charts. In qualitative research, documents are not “background”; they are potential primary data when your questions concern texts, practices encoded in artifacts, or institutional meaning-making.

Types of documents

• Public records: laws, reports, websites—often accessible but require contextual interpretation.
• Personal documents: diaries, letters, portfolios—high intimacy; consent and harm considerations intensify.
• Organizational documents: SOPs, training decks, KPI dashboards—reveal formal and informal priorities when read critically.
• Media and digital traces: news, forums, apps—raise provenance, bot, and archival volatility issues.

Classify each document by authorship, audience, purpose, and genre before heavy interpretation.

Authentication and contextualization

Ask: Who produced this, when, under what constraints, and for whom? Triangulate with interviews or observations when documents conflict with practice (“official story” vs “work-as-done”). Note redactions, version control gaps, and translation effects if documents cross languages.

Content analysis basics (qualitative orientation)

Qualitative content analysis codes meaning units into categories while preserving context. Steps typically include: define corpus boundaries; sample or include all relevant texts; define recording units (paragraph, post, case note segment); develop codes inductively, deductively, or hybrid; maintain a codebook with examples; iterate definitions as patterns clarify.

Discourse analysis basics

Discourse-oriented reading examines how language constructs subjects, objects, and power relations. Useful prompts: What is presupposed? Who is included/excluded? What metaphors recur? What silences matter? Align discourse analysis commitments (e.g., critical, Foucauldian, discursive psychology) with your paradigm statement.

Coding documents in a grounded theory framework

Treat a document segment as an incident: compare segment to segment and segment to emerging concepts. Avoid letting document headings dictate your categories if the data pushes elsewhere. Keep a memo trail for interpretive leaps (e.g., why a policy clause signals “risk shifting” as a category).

“All is data” and critical reading

GT’s “all is data” does not mean naive acceptance. Documents can misrepresent, lag reality, or perform legitimacy. Critical reading asks what the text does socially while still mining it for participant/informant meanings embedded in wording and format.

Practical workflow

1. Inventory documents with metadata (source, date, access path).
2. Select chunks relevant to the evolving theory; avoid infinite corpus sprawl.
3. Apply consistent naming for files and excerpts used in publications.
4. Link excerpts to audit trail entries (why this passage matters analytically).

Checklist

• Corpus boundaries and inclusion criteria documented.
• Provenance, authorship, and audience analyzed.
• Coding approach matches methodology (GT openness vs structured content analysis).
• Potential bias/misrepresentation flagged and triangulated where possible.
• Confidentiality and copyright constraints respected.

References (starting points)

• Bowen, G. A. (2009). Document analysis as a qualitative research method. Qualitative Research Journal.
• Prior, L. Using documents in social research.
• Glaser, B. G. Doing Grounded Theory — documents as comparable incidents.

Quick reference

Before coding a page: note genre, intended audience, and production constraints (who could edit this, and why it exists).

While coding: treat each excerpt as an incident; compare to interviews/observations when interpretations hinge on “official vs practiced” gaps.

In write-ups: distinguish textual claims from your interpretation; flag uncertain authorship or missing pages rather than smoothing gaps silently.