Reviewed and updated: 21 July 2026.
A research memo records what the researcher thinks the data may mean while keeping interpretation connected to evidence. It is not an AI summary of everything the participant said.
Quick verdict: correct material transcript errors, write the memo soon after the interview, separate evidence from inference, record contradictions and alternative explanations and link every emerging claim back to source data.
Commercial disclosure: NERALVO sells the Halo AI voice recorder and DOWAY transcription workflow.
What a research memo does
A useful memo can:
- Capture early interpretation
- Preserve context
- Record surprises
- Identify contradictions
- Compare cases
- Refine future questions
- Make analytical decisions auditable
Correct before interpreting
Check:
- Speaker labels
- Names or participant references
- Technical terms
- Numbers and dates
- Negative wording
- Sections marked unclear
- Translation choices
Preserve uncertainty where the audio cannot resolve the wording.
Write the first memo soon after the interview
Context fades quickly. Record observations about setting, rapport, interruptions, emotion and non-verbal context where the approved methodology permits it. Distinguish these observations from participant statements.
Use a consistent memo structure
- Interview reference and context
- Immediate overall impression
- Strong themes
- Contradictions or surprises
- Illustrative extracts
- Researcher interpretation
- Alternative explanations
- Questions for future interviews
- Links to codes or prior cases
- Methodological reflection
Separate evidence from inference
| Layer | Example |
|---|---|
| Evidence | Participant described repeating the same task each evening |
| Interpretation | The process may create hidden administrative burden |
| Alternative explanation | The issue may reflect local training rather than system design |
| Question | Do other participants report the same repetition? |
Use source-linked extracts
For every important quotation or observation, include the transcript line, timestamp or coded source reference. Avoid copying very long passages into the memo when a concise extract and link will suffice.
Record negative cases
Do not only collect data supporting the emerging theory. Mark participants, statements or events that conflict with it. Negative cases can reveal conditions, subgroups or weaknesses in the analysis.
Compare across interviews
After several memos, create comparison fields:
- Theme present or absent
- Strength of evidence
- Context
- Contradiction
- Participant group
- New question
Record changes to the interview guide
If a memo suggests new questions, document why the guide changed, when the change started and which participants received the new wording.
Use AI carefully in analysis
AI can suggest clusters or questions, but researchers should check:
- Whether important minority views were removed
- Whether frequency was mistaken for importance
- Whether a theme was invented
- Whether context was flattened
- Whether sensitive identifiers were exposed
- Whether the model's output is reproducible enough for the method
Preserve reflexivity
Record how the researcher's role, assumptions, interests and relationship with participants may affect interpretation. AI output does not remove researcher influence.
Protect participant identity
Use approved participant IDs, remove unnecessary identifiers and store re-identification keys separately. Quotations, rare events and locations may reveal identity even without names.
Use a memo header
- Study
- Participant ID
- Interview date
- Researcher
- Transcript version
- Memo version
- Confidentiality level
- Related codes
Review memos during analysis
Memos should evolve. Add dated entries when later interviews support, weaken or change the interpretation rather than silently rewriting the original thought.
Using NERALVO Halo for approved research
NERALVO Halo provides NOTE mode, CALL mode for supported and permitted calls, 64GB local storage, up to 35 hours of recording and Bluetooth connection to DOWAY. DOWAY supports transcripts, summaries, templates, translations, mind maps and exports, with one year of DOWAY Max included.
Researchers should use the complete Halo and DOWAY workflow only where approved by the protocol and institution. The researcher remains responsible for interpretation and auditability.
Research-memo checklist
- Material transcript errors corrected
- Memo written promptly
- Context recorded
- Evidence separated from inference
- Alternative explanations included
- Negative cases preserved
- Extracts linked to source
- Future questions recorded
- Researcher reflexivity included
- Participant identity protected
- Versions and changes auditable
Frequently asked questions
Can AI write the research memo?
It can propose a structure or questions, but the researcher must make and document the interpretation.
Should every theme be included?
Prioritise relevance while recording anomalies and minority views that may challenge the emerging analysis.
When should memos be written?
Soon after each interview, then revisited across the dataset.
Is a memo part of the research data?
It may contain personal or sensitive information and should be governed under the approved research-data plan.
Think with the data, not instead of it
A strong memo makes the path from participant speech to analytical finding visible, contestable and useful.
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