Quantitative research
Surveys, experiments, administrative data, and statistical analysis where measurement, comparison, and inference are central.
Design principle
Research begins with the question and the decision the finding is intended to support—not with the instrument. Concepts, population, time, and geography are reviewed before selecting design, data sources, sample, instrument, and analysis plan.
Research cycle
- Define the problem, purpose, and intended users.
- Formulate questions, objectives, concepts, and indicators.
- Select design, data sources, and sample.
- Develop, translate, review, and test instruments.
- Prepare fieldwork, training, quality, and ethics plans.
- Collect data, monitor implementation, and document deviations.
- Clean, code, weight, and analyze data.
- Review findings, tables, narrative, and limitations.
- Prepare deliverables and methodological disclosure.
- Publish or deliver and manage corrections and data.
Possible methods
Qualitative research
Interviews, focus groups, observation, and document analysis to understand experience, meaning, context, and mechanisms.
Mixed methods
Structured integration of measurement and explanation, identifying each component’s role and how results are combined.
Secondary research
Analysis of existing data, documents, and sources with assessment of quality, compatibility, and definitions.
Participatory research
Involve relevant groups in questions or interpretation without transferring professional responsibility to them.
Digital research
Use digital tools or data subject to privacy, quality, and transparency controls.
Sampling and inference
- Define population, frame, and unit of selection.
- Determine size and allocation according to precision, comparison needs, and resources.
- Document exclusions, nonresponse, and replacement.
- Apply and describe weighting where needed.
- Do not call a sample representative unless the design supports it.
- Use caution with small cells and multiple comparisons.
Instrument design
- Clear, neutral language appropriate to participants.
- One concept per item where possible.
- Balanced and reasonably exhaustive response options.
- Ordering that reduces leading and context effects.
- Logical, technical, and cognitive testing.
- Version, translation, and change documentation.
Data collection and quality
Quality controls differ by collection mode and may include training, testing, observation, verification calls, duration and logic checks, duplicate and location checks, authorized audio review, and open-response review. Automated indicators are not used alone for punitive decisions without human review.
Analysis and interpretation
- Analysis plan tied to research questions.
- Clear indicators, rules, and missing-value treatment.
- Attention to sample design and uncertainty.
- Distinction among description, association, and causation.
- Alternative specifications and sensitivity analysis where needed.
- Presentation of meaningful differences without selection.
- Interpretation connected to context and limitations.
Related standards
Quality Assurance
Review, responsibilities, checks, and error treatment.
Research Ethics
Consent, voluntariness, harm minimization, and safeguarding.
Privacy and Data Protection
Data minimization, security, retention, and sharing.
Independence and Transparency
Limits on commissioner influence, funding, and conflicts.
Publication and Disclosure
Information accompanying each output and methodology.
Corrections and Withdrawal
Treatment of errors, revisions, and invalid materials.
Digital Research and AI
Controls for digital tools, automation, and human verification.