Reviewed and fact-checked: 21 July 2026.
Focus groups produce insight through interaction. Participants agree, challenge, remember examples and reshape their answers in response to the group. That richness cannot be reduced safely to an automatic list of themes without preserving context, speaker uncertainty and the research method.
Quick verdict: approve the recording and data workflow before recruitment, explain it clearly to participants, design the room for attribution, correct the transcript before analysis and keep a visible chain from source audio to coding and final claim.
Commercial disclosure: NERALVO sells the Halo AI voice recorder. This guide describes a general research workflow and does not replace an organisation’s ethics, legal, information-security or data-protection requirements.
Start with the research question, not the recorder
Before collecting audio, define:
- The research question and why group interaction is useful.
- Who should take part and who should not be grouped together.
- Whether identifiable audio is necessary.
- Which transcription or AI services are approved.
- Who will access raw audio, transcripts and coding files.
- How quotations will be anonymised or attributed.
- What withdrawal arrangements apply.
- How long each record type will be retained.
The easiest recording method is not automatically the most appropriate research method.
Explain the full data journey before participation
Participant information should describe, in accessible language:
- That the session will be recorded.
- Why audio is required.
- Whether automated transcription or other AI processing will be used.
- Where processing occurs and which approved provider is involved.
- Who can hear the audio or read the transcript.
- How names, voices and quotations may be used.
- How confidentiality is limited in a group setting.
- How long files will be retained.
- How questions, complaints or withdrawal requests are handled.
Do not introduce an external processing service after the session without checking whether the approved participant information and research governance permit it.
Confidentiality in a group has limits
The researcher can control their own storage and publication workflow, but cannot guarantee that participants will forget or never repeat what others said. Set clear ground rules, avoid grouping people where disclosure could create harm and remind participants not to share identifiable contributions outside the session.
Design the room for usable attribution
| Control | Purpose |
|---|---|
| Manageable group size | Reduces overlap and improves meaningful participation |
| Consistent seating | Supports a speaker map |
| Similar recorder distance | Reduces large volume differences |
| Quiet, soft-furnished room | Reduces noise and echo |
| Clear participant identifiers | Supports accurate correction and anonymisation |
| Observer field notes | Capture non-verbal context and speaker order |
Use a moderator and an observer where possible
The moderator should focus on participation, questions and safety. An observer can maintain:
- A seating and participant-code map.
- Attendance changes.
- Non-verbal agreement, discomfort or laughter where analytically relevant.
- Major overlaps and interruptions.
- References to materials, images or activities not captured in audio.
- Immediate safeguarding or distress concerns according to the approved protocol.
Field notes should supplement the recording, not silently replace missing evidence.
How NERALVO Halo may fit approved small-group research
NERALVO Halo is an ultra-slim, phone-mounted AI voice recorder with NOTE mode for suitable room sessions, 64GB local storage, up to 35 hours of recording and Bluetooth sync with the DOWAY app. DOWAY can generate transcripts, summaries, templates, translations, mind maps and exports. One year of DOWAY Max is included.
Local capture may be useful where connectivity is limited, but later processing must still be approved for the research and participant data involved. Halo is not a substitute for specialist multi-microphone equipment when accurate attribution is critical.
Make the session easier to transcribe
- Record clear introductions or participant codes in sequence.
- Ask one question at a time.
- Use participants’ codes when directing follow-ups.
- Allow silence rather than filling every pause.
- Stop persistent side conversations.
- Ask for a material point to be repeated after overlap.
- Spell names, places and specialist terms where relevant.
- Announce activities, stimulus changes and breaks.
Separate the source transcript from the research transcript
A raw automated transcript is a machine-generated working draft. A research-ready transcript may require:
- Corrected speaker labels.
- Standardised participant codes.
- Verified names, figures and terminology.
- Marked overlap, laughter, silence or emphasis where analytically relevant.
- Removal or masking of identifying details.
- Clear uncertainty markers.
- Links to field-note references.
- A record of who corrected the transcript and when.
Use honest uncertainty
| Problem | Safer notation |
|---|---|
| Words cannot be heard | [inaudible 00:24:11] |
| Speaker uncertain | Participant uncertain |
| Two interpretations possible | [unclear: cost or course] |
| Several people overlap | [overlapping discussion] |
| Identifying detail removed | [workplace removed] |
A plausible invented phrase is more damaging to the research trail than a visible gap.
An AI summary is not thematic analysis
Automated tools may suggest topics, but a defensible analysis requires a documented method. Researchers should define:
- The unit of analysis.
- Whether coding is inductive, deductive or combined.
- How the codebook is developed and revised.
- How contradictory evidence is handled.
- Whether more than one researcher reviews coding.
- How themes relate back to the research question.
- How software-assisted suggestions are accepted, rejected or modified.
Frequency alone does not determine importance. A rare comment may reveal a serious risk, while repeated agreement may reflect group pressure.
Protect quotations from re-identification
Removing a name may not be enough. A quotation can identify someone through job title, location, unusual event, medical history or distinctive wording. Before publication:
- Check the quotation against the complete source context.
- Confirm it represents the intended theme.
- Remove unnecessary identifying detail.
- Assess whether the remaining combination still identifies the participant.
- Apply the approved attribution or anonymisation rule.
- Keep a controlled link to the source where auditability is required.
Maintain an auditable evidence chain
| Stage | Record |
|---|---|
| Recruitment | Approved participant information and eligibility process |
| Session | Audio, participant map and field notes |
| Transcription | Raw output, corrected transcript and correction log |
| Coding | Codebook, coding decisions and revisions |
| Analysis | Theme development, negative cases and researcher interpretation |
| Reporting | Claims linked to checked evidence and approved quotations |
| Closure | Retention, archive or deletion record |
Security and retention
Use approved accounts and controlled storage. Avoid leaving audio indefinitely on the device, in personal cloud folders or duplicated across email and chat. Set separate retention periods for contact details, source audio, identifiable transcripts, anonymised analysis files and final reports. Delete working copies when they no longer serve the approved purpose.
Focus-group checklist
| Before | During | After |
|---|---|---|
| Approve the research and processing plan | Confirm ground rules | Secure and inventory files |
| Prepare participant information | Use participant codes | Correct attribution |
| Test room coverage | Control overlap | Remove identifying detail |
| Create seating map | Maintain observer notes | Apply documented coding method |
| Set retention periods | Respond to distress or withdrawal | Link claims to checked evidence |
Frequently asked questions
Can AI perform the thematic analysis automatically?
It can suggest patterns or organise material, but researchers remain responsible for method, interpretation, contradictory evidence and final claims.
Can one compact recorder handle every focus group?
No. Group size, room acoustics and attribution requirements may justify specialist or multi-microphone capture.
Should raw audio be retained forever for auditability?
No. Retain each record only for the approved purpose and period.
Can participants be promised complete confidentiality?
The researcher can promise controlled handling, but other participants may repeat what they hear. Explain that limitation clearly.
What if a speaker label cannot be confirmed?
Mark it as uncertain or use a neutral participant label. Do not guess.
Keep the research trail visible
A reliable focus-group workflow connects participant information, source audio, corrected text, analytical decisions and final claims. AI can accelerate parts of the process, but accountability remains with the researcher.
Explore NERALVO Halo for approved small-group capture and structured transcript review.
Ready to capture meetings properly?
View the NERALVO Halo AI voice recorder with 64GB local storage, meeting capture, compatible phone-call recording workflows and one year of DOWAY Max included.
View NERALVO Halo