Why talk to people at all
Forms and counts show what people do or tick, but they say little about why. A conversation can uncover reasons, confusion and half-formed ideas that a checkbox never captures. A visitor might say they “liked” an interactive, and only in discussion explain that they never understood what the screen wanted them to do and tapped around until it moved.
Interviews and discussion groups are the main ways to gather this depth. They work with small numbers of people and produce words rather than percentages. That makes them good for understanding and poor for measuring how widespread an opinion is.
One-to-one interviews
An interview is a structured conversation between one interviewer and one visitor. It can be short, perhaps five to ten minutes at the exit of a gallery, or longer if the person agrees to schedule time.
Good interviews tend to share a few features:
- A short guide, not a script. Prepare a handful of open questions and follow the answers. “Tell me what you did at that screen” is stronger than “Did you find the screen easy to use?”
- Neutral wording. Avoid hinting at the answer you hope to hear.
- Comfortable silence. Waiting a few seconds after an answer often prompts people to add more.
- Follow-up prompts. “What made you say that?” or “Can you give an example?” draw out reasoning.
- Notes or recordings with permission. Tell the person what will be recorded and how it will be used. Visitor consent matters whenever audio or video is involved.
Interviews can happen while someone uses an exhibit, sometimes called a think-aloud session, where the visitor describes what they are doing and expecting. This is closely related to testing with rough versions, which is a form of prototype testing. Here the focus is on what people say about their experience.
Focus groups
A focus group brings together a small group, often somewhere between six and ten people, to discuss a topic led by a moderator. The discussion may be built around a set of questions, images, a rough design or a demonstration.
The format is useful when you want to hear how people talk to each other about a subject. One person’s remark can prompt another to recall something they would not have mentioned alone. Groups can also surface the shared language and assumptions of a community, which helps in early planning and in testing wording.
Typical uses include:
- Exploring how a community feels about a sensitive subject before content is written.
- Reacting to rough concepts or storyboards.
- Hearing how teachers, parents or other specific groups would use an exhibit.
- Following up on survey results that surprised the team.
The biases of small groups
Groups produce data, but it is skewed in predictable ways. Knowing these helps you read results with the right level of caution.
Dominant voices. One confident or opinionated participant can steer the conversation, and quieter people may agree to avoid conflict. A good moderator invites others in, but cannot remove the effect entirely.
Social pressure. People often moderate their views in front of strangers, particularly on contested or personal subjects. A group may sound more agreeable than its members actually are.
Groupthink. Early opinions can set the tone, and later speakers align with them.
Self-selection. People who volunteer for a discussion are usually more interested, more articulate and more available than the average visitor. Weekday daytime sessions can exclude working people. If the audience includes older adults or families with young children, scheduling and childcare affect who can attend.
Moderator influence. Tone, facial expressions and the way a question is asked can lead people. This applies to interviews as well.
Small numbers. Eight people cannot represent all visitors. If five of eight dislike a design, that is a prompt to look closer, not a statistic. It would be misleading to say “most visitors dislike it” on that basis.
Because of these issues, treat focus group results as hypotheses to check. If several different methods point the same way, confidence grows. If only the group said it, hold it loosely.
Choosing between the two
Interviews suit sensitive topics, individual experiences, and people who may not speak freely in front of others. They also suit accessibility feedback, where one person’s needs can be quite specific. A discussion with a visitor who uses screen readers or one who finds crowded rooms overwhelming is often richer one to one.
Groups suit topics where shared meaning matters, such as community views on how a story should be told. They take less moderator time per person, but the transcripts are harder to interpret.
Many projects use both. A few interviews first can reveal what to explore, and a group can then test how widely the ideas resonate within a particular audience.
Planning a session
A modest plan covers these points:
- Purpose. Write the two or three things you need to learn. If a question does not serve one of them, drop it.
- Recruitment. Decide who should be in the room and how you will find them. Mixing very different people can be productive, but status differences, such as staff alongside visitors, can quiet some participants. Keep staff who designed the exhibit out of the room, or at least out of the discussion.
- Location and access. Choose a quiet, accessible, comfortable space. Provide plain-language invitations and, where needed, interpretation or other support.
- Opening. Explain the purpose, say there are no right answers, and make clear that criticism is useful, not rude.
- Moderator. Ideally someone who is not on the design team, trained to stay neutral. A second person can take notes.
- Thanks. Many institutions offer a small gift or thank-you. Check your own rules and make sure it does not become pressure to say positive things.
Recording and analyzing what you hear
Record or write detailed notes soon after, while they are fresh. Quotes are valuable, but a handful of colorful ones can give a misleading impression. Read all of the material, then sort comments into themes: what confused people, what they enjoyed, what they expected, what they suggested. Note how many people raised each theme and whether it came from one voice repeated or several independently.
Separate what people said they would do from what you saw them do. People are not always accurate in predicting their own behavior. Combine the talk with visitor observation where possible.
What to do with the results
Share a brief summary with the design team, including the limitations: who took part, how they were recruited, and what you cannot conclude. Make a short list of changes to try, then check them with another round of testing. Feedback gathered this way is most useful when it leads to a specific decision, and when the team is open to hearing something it did not hope for.