Narrata
Home / Research Standards

Research quality and participant care

Evidence people can trust—and teams can use

Good research is more than a set of methods. It is a chain of careful decisions about the question, participants, data, interpretation and claims. These are the working standards we bring to that chain.

Discuss a research project View a sample deliverable

Research becomes credible when the strength of each claim matches the evidence behind it. We design projects around a real decision, make limitations visible and distinguish what we observed from what we infer.

Our approach is adapted to the context, audience and risk of each engagement. A quick concept test and a sensitive community study should not be run in exactly the same way. The project plan sets out the appropriate safeguards, roles and review points before data collection begins.

These standards describe our default working approach. Project-specific requirements, organisational policies and relevant legal or ethical review processes are agreed during scoping.

What clients should expect

  • A clear decision and research question
  • Methods that fit the question
  • Respectful, informed participation
  • Purposeful data collection
  • Traceable analysis and proportionate claims
  • Explicit limitations and practical next steps

Six standards behind the work

The details change by project. The responsibility to make sound, transparent choices does not.

01

Start with the decision

We clarify what must be decided, what is already known and how the evidence will be used before selecting a method.

02

Recruit for relevance

Recruitment criteria follow the research question. We consider who may be missing, what barriers affect participation and which differences matter.

03

Make participation informed

People should understand the purpose, activities, recording approach, data use, voluntary nature of participation and how to raise a concern.

04

Collect with restraint

We aim to gather only the information needed for the agreed purpose and define access, storage, retention and deletion arrangements during planning.

05

Keep claims close to evidence

Analysis looks for patterns, differences and counter-evidence. We separate observations, interpretation and recommendation rather than blending them together.

06

Show uncertainty

Reports state the sample, method and limitations. We avoid turning directional qualitative evidence into unsupported population claims.

Quality checks across the project

  1. ScopingAlign the research question, intended decision, audience, existing evidence, constraints and level of confidence required.
  2. Study designDocument the method, participant criteria, instruments, responsibilities, logistics, consent and data-handling approach.
  3. Pilot and fieldworkTest the discussion guide or tasks, watch for ambiguity and bias, and record relevant changes to the plan.
  4. AnalysisUse a consistent evidence trail, examine contradictory cases and review interpretations against the source material.
  5. ReportingConnect each major claim to evidence, distinguish severity or confidence and make limitations visible to decision-makers.
  6. ActionTranslate findings into prioritised choices without overstating what the study can prove.

Participant care in practice

  • Plain-language study information
  • Voluntary participation and withdrawal
  • Proportionate incentives where appropriate
  • Accessible and context-sensitive sessions
  • Care around sensitive topics and power dynamics
  • An agreed path for questions or concerns

Specific safeguards are strengthened when a study involves sensitive subjects, vulnerable groups or meaningful risk to participants.

Responsible use of AI in research

AI can support parts of a workflow, but it does not remove human accountability for participant care, interpretation or the final recommendation.

Protect source material

We do not place identifiable or sensitive participant material into general-purpose AI tools without an agreed, appropriate data-handling arrangement.

Keep humans accountable

Researchers remain responsible for checking context, ambiguity, contradictory evidence, bias and the validity of claims.

Use tools for a defined purpose

Any AI-supported step should have a clear purpose and review process rather than being applied automatically to every project.

Disclose material use

When AI materially influences analysis or a deliverable, its role and the human review applied should be made clear to the client.

Transparent about what the evidence can support

Ask us about the proposed sample, consent process, data handling, analysis or limitations. A credible research plan should make those answers clear before fieldwork begins.

Discuss your researchInspect the sample audit