A research team wants to know why students avoid their university’s counselling service. To save time they run three focus groups of eight students each. In all three, nobody shares a personal experience, and each group quickly agrees that “the service is good but students are busy”. A deeply private topic has been pushed into a public format.
Focus groups are powerful when you want to see people discuss, disagree and form views together. They are a poor choice when all you want is a batch of individual answers in a short time.
When to use them and when not to
| Good fit for a focus group | Choose individual interviews instead |
|---|---|
| Exploring shared norms and the language a group uses for an issue | Private or sensitive experiences: mental health, income, violence |
| Testing reactions to an idea, leaflet or planned programme | Detailed life histories in sequence |
| Preparing a questionnaire: what words people use for the problem | Participants with hierarchical relationships to each other |
| Watching views being negotiated and changed through talk | You need independent views not shaped by others |
Group size and number of groups
Five to eight participants is the usual recommendation. Fewer than five and the session drifts into a small group interview; more than eight and quiet people disappear while the moderator constantly cuts people off. For complex topics or expert participants, four to six often works better.
Plan the number of sessions by type of group, not by total. If you want to compare first-year and final-year students, and men and women, you have four types; each needs at least two sessions so that one unusual session is not mistaken for the pattern. That is eight sessions minimum. Stop when new sessions within a type stop producing new ideas.
Composition: similar within groups, different between them
People in one group should share characteristics relevant to the topic, so they feel free to speak. Teachers should not sit with their head teacher; patients should never sit with their clinicians. If you recruit through an organisation, ask who knows whom: a group of close friends talks easily, but much of what “everyone knows” never gets said.
The differences you want to analyse should sit between groups, not inside one. That lets you compare without making anyone guarded.
The question route: few questions, funnel order
A 90-minute session usually has room for only five to seven main questions. Order them as a funnel:
- Opening: easy, factual, answerable by everyone and not about the core topic (course, year of study).
- Introductory: “When you hear ‘university counselling’, what comes to mind first?”
- Transition: talk about the last time you or a friend needed someone to talk to.
- Key questions (two or three): what puts people off going, and what would bring them in.
- Ending: “Of everything we discussed, what matters most?” followed by a summary the group can confirm or correct.
Small activities deepen the talk: ranking cards by importance, rating a draft leaflet, or writing privately on paper before sharing. Writing first is especially useful for reducing the pull towards whoever spoke first.
Moderation: draw out the quiet, rein in the dominant
- Set ground rules at the start: no right or wrong answers, different views are exactly what the study needs, respect what others share.
- Dominant speakers: thank them specifically, then redirect: “Thanks, that point is clear. I’d like to hear from people who haven’t spoken yet.” Use eye contact and body language to turn attention elsewhere.
- Quiet participants: invite them by name with an easy question, or ask whether their experience has been different.
- Fast consensus: ask the opposite question: “Has anyone seen it work the other way?” Quick agreement is often a sign of group pressure rather than truth.
- Stay out of the debate: the moderator never shares an opinion or nods approval at one side.
Use two people: a moderator and a note-taker. The note-taker tracks who speaks, in what order, non-verbal reactions and moments of silence or laughter. The recording will not capture these, and transcribing eight voices without notes makes it very hard to tell who is speaking.
Confidentiality: what you cannot promise
You can promise that the research team will keep things confidential, but you cannot guarantee that other participants will. Say so in the consent form, repeat it before starting, and advise participants not to share anything they would not want others in the room to know. This is one more reason to keep highly personal topics out of focus groups.
Analysis: the data are the conversation
A common analytic error is quoting individuals as if they had given eight separate interviews. The primary unit of analysis is the group. When coding, look for:
- Which views the group picks up and which it contests or ignores.
- Who changes position after hearing others, and why.
- Shared words and examples the group uses for the issue.
- Topics the group avoids or moves away from quickly.
In your write-up, a short exchange between two or three speakers is often more convincing than a single quote. Label group and speaker clearly, for example (Group 3, female, final year, G3-P5).
Logistics checklist
- A quiet room with a round or U-shaped table and no head seat.
- Two recorders at opposite ends of the table, batteries checked.
- Name cards with first names or pseudonyms only.
- Light refreshments and no more than 90 minutes.
- A consent form that explains the limits of confidentiality within the group.
- A debrief summary written immediately after, before moderator and note-taker part ways.
Next step: write down your research question and ask yourself whether you need to see people talk to each other or to hear each person’s own story. If it is the second, switch to individual interviews before you recruit a single participant.
Câu hỏi thường gặp
How many people should be in a focus group?
Usually five to eight per group. For complex topics or expert participants, four to six gives each person enough time to contribute.
How many focus groups do I need?
Plan by the number of group types you want to compare, with at least two sessions per type. Stop when new sessions in a type no longer add ideas relevant to your research question.
What is the difference between a focus group and a group interview?
In a focus group participants talk to one another and that interaction is the data. A group interview is often just an interviewer questioning each person in turn in the same room.
Are focus groups suitable for sensitive topics?
Usually not, because people hesitate to share in front of others and the research team cannot guarantee that fellow participants will keep things confidential. Individual interviews are generally a better fit for sensitive subjects.
How do you handle a dominant participant in a focus group?
Acknowledge their point specifically, then invite people who have not yet spoken. The moderator can use eye contact and gestures to redirect attention and remind the group that many perspectives are needed.
