Running the conversation
Focus groups and interviews that yield usable data
Preparation, ground rules, prompts, silence, dominant voices, interpretation and closing well.
A focus group is not a meeting, and an interview is not a chat. Both are structured conversations that produce a record someone else can analyse. This page covers the practical skill of running them, including when you co-facilitate and when you should not lead at all.
Before the session
The preparation checklist
- Read the topic guide, and know which three questions matter most.
- Check the practical details: venue, access, seating, water, food, toilets, interpreter, recording equipment, consent forms, payment.
- Confirm the safeguarding and escalation routes for that day.
- Test the recorder. Have a paper backup.
- Agree who does what, if you co-facilitate. Who opens, who asks, who watches the room, who takes notes.
- Arrive early, and be at the door when people arrive.
The opening, in five minutes
- Welcome people by name where you can.
- Say who you are, and why you are doing this work.
- Say what will happen, and how long it will take.
- Set the ground rules: no right answers, everyone's view matters, we will not use names in the report, you can leave or skip a question at any time.
- State the limits of confidentiality, plainly, before anything is said.
Asking questions that open a conversation
Question shapes, with examples
- Open: "Tell me about the last time you tried to get an appointment."
- Follow the word: "You said it was humiliating. What made it humiliating?"
- Concrete before abstract: ask about a specific occasion, not about opinions in general.
- Contrast: "What would have made that easier?"
- Summary check: "So the delay was less about the wait and more about not knowing what was happening. Have I got that right?"
Avoid questions that suggest the answer, questions with two parts, and questions that ask people to confirm your own view.
The skills that matter most
Six behaviours
- Silence. After a person stops, wait three seconds. The second sentence is often the honest one.
- Not filling the gap. If the room is quiet, ask again more simply rather than answering for them.
- Even airtime. Notice who has not spoken, and invite without singling out: "Is there a view we have not heard yet?"
- Managing a dominant voice. "That is a useful point. Let me bring others in, and come back to you."
- Staying neutral. You may show warmth. You do not argue, correct or educate during data collection.
- Closing well. Summarise, thank people specifically, explain what happens next, and say when they will hear the findings.
Co-facilitation
In the IBISES projects, focus groups were jointly facilitated by link workers and academic team members. That arrangement is the best of both: community legitimacy and methodological competence in the same room. It needs an agreement, made beforehand.
- Who leads the introduction, and who explains consent?
- Who holds the topic guide, and who watches the room?
- How will you signal to each other that a theme needs following or leaving?
- Who handles a distressed participant?
- Who debriefs with whom, and when?
Language and interpretation
The IBISES projects ran sessions with real-time English to Cantonese translation, minimal Urdu support, and sessions entirely in English. Three lessons.
- Decide the language of the session in advance, and tell participants.
- An interpreter adds time. Plan for a longer session, not the same session with a shorter discussion.
- Record the interpretation arrangement in the study record, because analysis depends on knowing what language the words were spoken in.
Inclusive communication, especially with aphasia or cognitive difficulty
Supported conversation, in practice
- Reduce background noise. One conversation at a time.
- Use short sentences, one idea in each.
- Give extra time. Do not finish sentences unless invited.
- Support with writing, pictures, gesture, or a communication chart.
- Check understanding by asking the person to tell you what they understood, not by asking "do you understand".
- Accept any way of answering: speech, writing, gesture, pointing, a familiar person's help if the person wants it.
- Never speak about the person to their companion as if they were absent.
When a session goes wrong
Five situations and the response
- A participant becomes distressed. Stop the questions. Acknowledge. Offer a break, water and privacy. Follow the study's distress protocol. Record it after.
- A participant discloses harm. Follow the safeguarding route in the ethics page. Tell them what you will do, and do it.
- One person dominates. Use the room, not the person: pairs, rounds, or a written pause.
- The recording fails. Say so honestly at the end, write up your notes immediately, and report it to the lead the same day.
- A conflict opens between community positions. Do not resolve it. Name it, record it, and let it stand as data.
Debrief
Debrief with your co-facilitator within 24 hours, while the memory is fresh. Cover what worked, what you would change, what you noticed in the room that is not in the recording, and what needs to go to the team. The IBISES model used debrief after every activity, with reflective notes from link workers and researchers. Those notes became part of how the model improved.
The test of understanding
You are co-facilitating a group of eight. One man has answered every question, and two women have said nothing. The guide has four questions left and time is short. What do you do?
Change the format, not the person. Move to a round, or to pairs, or ask people to write one line each before speaking. Then invite: "Let's hear a view we have not had yet." Protect the quiet voices without embarrassing them, and do not skip the remaining questions to manage the room.