Analysis & synthesisDocument analysisCoding frameInter-coder reliabilityOpen responsesQuantitative and qualitative
Content Analysis
Content Analysis
Also known as: Content analysis approach、Document analysis、Quantitative content analysis
Classifies text or records into common categories to examine concepts' presence, frequency and relationships. Specify units and code definitions and link findings to originals. [1]
- Difficulty
- Intermediate
- Cost
- Low
- Duration
- Planning example, not a universal standard duration: 1–3 weeks for frame development, coder training, implementation and reliability checks
Content updated: 2026-10-07Source checked: 2026-09-13 (includes AI cross-checking)No documented human reviewIncludes fictional examples
How it works at a glance
A concept can be counted as present at least once in a document or by its total occurrences. Define units and classification rules according to the question. [1]
Record included and excluded examples and ambiguous-case handling. Check agreement where several people classify under common rules, but no universal coefficient threshold guarantees semantic validity. [1]
- Confirm purpose and research questions
- Develop a coding frame
- Train coders
- Code the full dataset
- Calculate inter-coder reliability
- Aggregate, analyze and interpret
When to use it and when it is unsuitable
When to use it
- When systematically categorizing many open responses, documents or reports to examine frequencies and trends
- When several coders must share analysis and reproducibility is required
- When comparing policy documents, minutes or media reports over time or across actors
- When emphasizing systematic procedures and reproducibility over thematic analysis and needing numerical coding agreement
When it is unsuitable
- When estimating population proportions from few documents. Content can still be described with limited documents; distinguish purpose and generalization scope.
- When deeply exploring data before developing a coding frame. Open-ended thematic analysis is more appropriate.
- When audio or video dominates and transcription is prohibitively costly or slow
Participants, roles, and deliverables
Participants and roles
1 analysis lead and 1–3 coders; multiple coders are ideal for reliability checks. For large datasets, assign dedicated coder training staff.
Main deliverables
- Codebook: definitions, criteria and typical examples
- Inter-coder reliability: Cohen's kappa / Krippendorff's alpha
- Category counts and percentages for quantitative content analysis
- Content analysis report: methods, reliability, findings and interpretation
Implementation steps
- Confirm purpose and research questionsClarify what to count or classify. Choose quantitative content analysis for frequency or qualitative content analysis for interpretation.
- Develop a coding framePre-code some material and combine deductive categories from theory or questions with inductive categories from data. Document category definitions, decision criteria and typical examples in a codebook.
- Train codersWhere multiple coders are involved, share the codebook and practice on training data. Discuss disagreements and revise definitions. Skipping this stage substantially reduces reliability.
- Code the full datasetCoders independently code all data. Consistently choose either discussion-based consensus coding or comparison after independent coding.
- Calculate inter-coder reliabilityFor quantitative coding, examine independent agreement using kappa, alpha or other measures suited to scales and missingness. Report coefficients, units, category imbalance and disagreements. Do not judge every analysis by one threshold.
- Aggregate, analyze and interpretOrganize counts, percentages and denominators for quantitative analysis, or meanings and relationships for qualitative analysis. Explain reliability methods and limitations rather than requiring the same coefficient everywhere.
Strengths and limitations
Strengths
- Explicit definitions, coding rules and independent checking make analytical processes easier to assess.
- Systematically and efficiently processes large text datasets and integrates with quantitative data.
- Enables secondary analysis of existing documents and records without new collection.
- Quantitative findings can support statistical tests and hypothesis-testing analysis.
Limitations
- Attention to surface meanings makes context, implications, irony and hidden intentions harder to capture.
- Frame quality shapes the whole analysis, requiring substantial development time and expertise.
- Large-scale coding is labor-intensive; text mining becomes a practical complement beyond several hundred cases.
- High inter-coder reliability indicates agreement, not whether codes accurately capture meaning.
Common pitfalls
- Arbitrary coding due to inadequate definitions. An ambiguous category such as statements about difficulties produces differing coder judgments. Specify criteria and typical and atypical examples.
- Not reporting reliability evidence. For independent agreement, state coefficient calculation conditions and interpretation. For dialogue or consensus analysis, document deliberation and judgments; multiple analysts alone do not demonstrate quality.
- Non-exclusive categories complicate aggregation when 1 text unit can receive multiple codes. Refine coding units and category structure beforehand.
Conditions of use and evidence
- Distinguish concept presence from word occurrences, and check category assignment rules and context. [1]
Classify application barriers
A fictional illustrative example where two people classify 10 open responses about applying for a consultation event.
Scroll horizontally to view the table
| Fictional response | Code | Reason |
|---|---|---|
| The screen went back during entry. | Operational barrier | Interrupted application operation |
| The screen was easy to use. | No operational barrier | Mention of operation alone is not a barrier. |
| Too busy to apply | Time barrier | Distinct from operating difficulty |
Common errors in judgment
Common mistake
Assigning the same category whenever the same word appears
How to revise
Set contextual and exclusion rules.
ReasonThe same word can occur in negative statements or with different meanings.
Common mistake
Assuming agreement because two coders' totals match
How to revise
Check whether identical units received identical codes.
ReasonTotals may match despite different cases being classified.
Organize your thinking by writing
Content analysis codebook and reliability worksheet
A practical template is available to fill in and print on A4 paper.
Open worksheetRelated methods
References and checked sections
- Content AnalysisColumbia University Mailman School of Public Health
- Section consulted
- Conceptual Analysis / Reliability and Validity
- How it relates to the text
- Distinguish concept presence from word occurrences, and check category assignment rules and context.
Editorial and checking information
How to cite this page
EvalCompass Editorial Team. “Content Analysis”. EvalCompass, updated 2026-10-07. https://evalcompass.jun-nakatani.com/en/methods/content-analysis/
Add the date you accessed the page. Text created by this site is available under CC BY 4.0. Quotations, external materials, and bundled fonts have their own terms of use. When reusing content, credit the source, link to the license, and indicate any changes.
Change history
- — Checked sources for definitions and applicability conditions, and added fictional examples, reasons for judgments and examples for every worksheet field. Revised unverified claims about history and adoption. Also reviewed existing procedures and cautions, and adjusted fixed participant counts and thresholds and generalizations about institutions. Preserved field positions, types and counts while revising fixed conditions and thresholds in the forms.
- — Correction based on independent content review (content-check-causal). Replaced mandatory reliability coefficients with reporting evidence appropriate to the analytical approach.