How to Transcribe a Focus Group

A focus group needs more than an automatic transcript: record it with clear consent, make the speakers as easy to hear as possible, then use the transcript as a first draft that you label and verify before analysing. The difficult part is not turning sound into words; it is preserving who said what when people respond to each other.

Start with consent and a recording plan

Tell participants that the session is being recorded, how the recording and transcript will be used, who can access them and how long they will be kept. The exact consent rules depend on where you work and the type of research, so follow your organisation's approved process rather than relying on a generic template.

Give the moderator a short opening script with the consent question and record the answer. It is easier to keep that evidence with the session than to reconstruct it later.

Capture cleaner voices

• Ask everyone to say their name once at the start; it gives you an audio reference for later labels.
• Put the microphone close to the group, away from a projector, fan or open window.
• Ask participants to avoid talking over one another when possible, while accepting that natural discussion will sometimes overlap.
• Keep a written seating plan or participant list so a speaker label can be confirmed later.

One microphone in a lively room is a compromise. It can produce a useful transcript, but it cannot reliably separate simultaneous speech. Separate microphones or individual tracks are better when the identity of every speaker matters.

Transcribe, label and review

1. Upload the recording to audio to text and select the spoken language when you know it.
2. Use the speaker labels as a first pass, then rename and correct them from your participant list.
3. Listen again at overlaps, unclear answers, names and statements you expect to quote.
4. Remove or replace identifying details in the working copy if your research plan requires it.

Prepare text for analysis

Decide before you edit whether your project needs a clean transcript or a verbatim one. A clean version is easier to read and often enough for thematic analysis. A verbatim version keeps filler words, false starts and interruptions, which can be important when the way something was said is part of the analysis.

Keep timestamps beside passages you may quote. They are the quickest way to revisit tone, a pause or a response from the rest of the group that plain text cannot show.

Do not over-trust speaker labels

Speaker recognition is helpful for making a long discussion readable, not a final identity check. Similar voices, people moving around the room and crosstalk can all swap labels. Verify any label attached to a quote, finding or participant-level conclusion before using it.