Use Word when you need to correct an interview transcript into a readable document. Use NVivo when you need to organise and analyse passages across interviews. Consider Notibo when permitted audio needs a working transcript and a concise note for your own review. These tasks can sit in the same dissertation, but they are not interchangeable. Your approved handling of participants' information comes before the feature comparison.
Notibo publishes this guide. The comparisons below use Microsoft, Lumivero and university documentation checked on 15 October 2026; they are not an accuracy benchmark. For the wider timetable around interviews, see our UK dissertation guide.
What should your interview produce?
Define the object you need before selecting the software. A transcript represents what was said. A research memo records your developing interpretation. A coding system groups evidence according to the method you have chosen. A summary is a compressed account that may help you orient yourself, but it cannot stand in for all three.
Suppose your supervisor asks how a participant qualified a claim about working during term. Sending a polished paragraph about student employment may not answer the question. You need the relevant passage, the surrounding exchange and the wording that limits the claim. If several participants discuss the same issue, you may then need a consistent way to compare them.
| Work to complete | Word | NVivo | Notibo |
|---|---|---|---|
| Correct speech recognition | Transcript editor with associated audio in supported Transcribe accounts | Prepare or review source material within the chosen import workflow | Edit generated text, checking against a separately retained source where needed |
| Write an interview memo | Write and format the memo yourself | Link interpretation to an organised research project | Generate a starting summary and revise it |
| Apply a codebook across interviews | Possible manually, but you manage the structure | Dedicated codes and reporting tools | Not an equivalent qualitative coding environment |
| Hand over a readable document | Native Word document | Export relevant material or codebook | Export notes or text to Word/PDF |
| Preserve research evidence | Use approved storage and a deliberate naming scheme | Manage project files and linked media carefully | Do not treat the temporary audio copy as your research archive |
The final row matters even when the output looks excellent. Losing the source can prevent you from resolving a later disagreement about wording.
When is Word the practical first step?
Microsoft's Transcribe documentation describes direct recording or audio upload, speaker-separated text, correction alongside timestamped playback and adding transcript text to a document. Availability depends on the Microsoft account and supported application. Check the actual account menu rather than assuming every university licence includes the same service.
Word is a useful candidate when your immediate deliverable is a corrected transcript that your supervisor can read and comment on. You can keep the speaker naming and formatting convention consistent with your methods chapter. A transcript is still an interpretation of audio, so decide how to represent pauses, interruptions and uncertain words with your supervisor.
Try one permitted practice file before importing a series of interviews. Find a speaker transition, correct a technical name and add the relevant section to a document. Open that document outside the transcription pane. Check which context remains available to its recipient and which still depends on the audio file and permissions.
What does NVivo add after transcription?
NVivo's value is the research organisation that follows source preparation. Lumivero's documentation for exporting codes in NVivo 15 describes code lists, references and codebook exports. Those are tools for documenting an analysis, not evidence that an automated summary has already performed it correctly.
The University of York's qualitative-data guide discusses several ways to work with text and audio, including direct audio annotation and preparing transcript styles for NVivo. This is a useful reminder that a full automatic transcript is one possible starting point. Your method may call for a different level of transcription or direct engagement with the recording.
For a small undergraduate project, a carefully organised set of documents may be sufficient if your supervisor agrees. For multiple interviews with a developing coding structure, a dedicated research project can make decisions easier to trace. Choose based on the analysis you will actually perform, not the apparent sophistication of the application's name.
Why does the university account matter in the UK?
The university's permission to use a desktop application does not necessarily extend to every associated cloud feature. York's guide specifically tells researchers to check whether transcription services are approved for their data. Apply that question to Word, NVivo's additional services and Notibo alike.
Write down the proposed route: where the original recording is stored, which service receives it, who can access the transcript and when each copy is removed. Confirm that route through your supervisor or the relevant university process. A participant's name removed from a filename does not show that their voice or story has been anonymised.
Do not upload real interview material merely to decide which interface you like. Make a practice recording with no participant data. Our UK study-data guide provides broader background, while the project-specific decision remains with your institution.
How do you compare outputs without rewarding a misleading summary?
Consider this invented exchange: a participant says that recorded lectures helped during two missed sessions, but that they would still choose to attend a discussion class. A weak summary might reduce that to a general preference for recorded teaching. The words sound plausible while the qualification has disappeared.
Keep three columns in your evaluation notes: the checked statement, the generated output and the correction needed. A useful transcript should preserve the qualification. A useful review memo should distinguish the participant's account from your interpretation. A useful code should have a definition that you can apply consistently when another interview presents a different experience.
Test a quotation and a disagreement
Choose one sentence you might quote and one passage that complicates your emerging view. Find both in the original. Check whether any tool silently joined separate answers, removed a negation or presented an interviewer suggestion as a participant claim. The purpose is to expose the errors that would change your analysis, rather than compare punctuation preferences.
Check the handover
Export the working document, close the application and reopen the copy from the location you intend to use. Confirm the file name, participant code and version. If your workflow depends on linked audio, test that link through the approved account. This small rehearsal can reveal more than processing all interviews immediately.
Where does Notibo fit?
Notibo transcription can help you orient yourself in audio you are authorised to process. Its summary is a working aid for reading and review, and its text remains editable. It does not replace a checked quotation, a documented codebook or your own analytical judgement.
Save the source file if your approved workflow requires it. Notibo normally removes source audio after successful processing; retained audio following a failure or quality recovery is temporary, up to 24 hours. Automatic quality checks do not certify that every sentence is correct.
For a dissertation, choose Word for the document-correction task and NVivo where structured analysis warrants it. Add Notibo only when its review notes solve a separate, permitted need. Compare the trial and current pricing using practice material before committing your research workflow.