3 Focus Groups, Observation, and Projective Techniques
Where this chapter fits. Step 4 of the research process is to design the method and the forms. Chapter 2 covered one method, the interview. This chapter covers the other main ways to collect qualitative data and how to choose among them.
3.1 A Library That Teenagers Did Not Use
A public library system rebuilt one branch with a room for teenagers. The case is a composite written for this book, and the quotations in it are synthetic.
A year later the room was nearly empty after school. The library had surveyed teenagers who held library cards. Most of them marked “satisfied.” The survey reached the teenagers who already came. It did not explain why the others stayed away.
The staff tried three other methods.
First, they held two focus groups with teenagers from the nearby high school. In a group, the students talked to one another more than to the adult in the room.
Synthetic
“It’s nice in there. It’s just that the desk is right there.”
“Yeah, they can see everything you do.”
“My brother got told to be quiet for laughing. One time. He never went back.”
Second, a staff member observed the branch on ten afternoons. She recorded where teenagers sat and how long they stayed. Most of them walked past the new room. They sat at two tables behind the tall shelves, out of sight of the desk.
Third, the staff asked 30 students to finish a few sentences in writing. One sentence began “Kids who hang out at the library are.” The most common endings were about having nowhere else to go.
Each method showed something the others could not. The groups showed that being watched was a shared complaint. The observation showed what teenagers did about it. The sentences showed a reputation that no student had raised aloud in front of others.
This chapter covers these three methods and a few related ones. It ends with a way to choose among them.
3.2 More Than One Way to Collect Words
Chapter 2 treated the one-to-one interview in detail. Researchers also call it a depth interview. It remains the most common method in qualitative research, and it is the right choice for a private topic or for one person’s full account.
An interview has two limits. It records what a person says, which can differ from what the person does. It also records only what a person is able and willing to say to an interviewer.
The other methods address those limits.
- A focus group adds the reactions of other people.
- Observation records what people do.
- A projective technique asks an indirect question when a direct one would fail.
- A case study combines several methods to study one setting in depth.
No method is best in general. Each fits some questions and fails on others.
3.3 Focus Groups
A focus group brings six to ten people together for a guided discussion of 60 to 90 minutes. A moderator asks the questions and manages the conversation. An assistant usually takes notes on who said what.
A focus group is more than a way to interview eight people at once. Its value is the interaction. Participants respond to one another, disagree, and build on what others say. One person’s story reminds another of her own. The library students did this in the exchange above.
When a Group Fits
A group suits a topic that people discuss with others in ordinary life. A neighborhood, a product, and a public service are examples. A group can show which views are shared and which are disputed. It can also show the words people use with one another, which helps a researcher who will later write a survey.
A group is a poor choice for a private topic. A parent may not describe a custody dispute in front of seven strangers. A group also fails when the researcher needs each person’s full account. Each participant speaks for only a few minutes in total.
Who Is in the Group
Participants in one group should be similar on the characteristics that affect how freely they will talk. Teenagers speak less when adults are in the group. Employees speak less when a supervisor is present. Researchers call the groups that result segments.
The comparison then takes place between groups. A study of a library might hold separate groups for teenagers who use the branch and teenagers who do not.
Strangers usually make a better group than friends. Friends share a history that the moderator cannot see, and they may avoid disagreeing with one another.
How Many Groups
One group is never enough. A single group can be shaped by one strong personality, and the researcher cannot tell a pattern from an accident. Plan at least two groups for each segment.
Research on real studies gives rough guidance. Guest, Namey, and McKenna (2017) analyzed 40 focus groups and tracked when each theme first appeared. Most themes had appeared within the first two or three groups, and nearly all within three to six. Many applied studies hold four to eight groups in total.
The Moderator’s Guide
A moderator needs a shorter guide than an interviewer does. Four to six main questions are enough for 90 minutes. Table 3.1 shows the usual order.
| Part | Purpose | Example for the library study |
|---|---|---|
| Opening | Every person speaks once, on an easy topic. | Tell us your first name and where you usually go after school. |
| Introductory | The topic is raised in general terms. | What comes to mind when you think about the library? |
| Key questions | The two to four questions the study depends on. | Think about the last time you went to the branch. What was that visit like? |
| Closing | Participants sum up and add what was missed. | If you could change one thing about the branch, what would it be? |
The rules for good questions in Chapter 2 apply here too. Ask open questions, ask one thing at a time, and do not suggest the answer.
What the Moderator Does
The moderator has three tasks beyond asking the questions.
- Draw out quiet participants. “We have not heard from this side of the table” is enough.
- Limit dominant participants. Thank the person, then turn to the others by name.
- Stay neutral. A nod at one opinion tells the group which answers are welcome.
The moderator also sets ground rules at the start. One person speaks at a time. There are no right answers. Disagreement is welcome.
What a Group Cannot Show
A focus group has three limits.
The group shapes what is said. Some participants agree with the room and keep their own view private. This is conformity. Others give the answer that sounds most acceptable. This is social desirability bias, and Chapter 5 of Before the Numbers describes it for surveys.
The group is not a sample of a population. Thirty students in four groups cannot show what share of all teenagers hold a view. A statement such as “most teenagers feel watched” needs a survey.
The unit of analysis is the group. Eight people in one room have influenced one another. Four groups provide four discussions to compare. They do not provide 32 separate opinions.
Online groups are now common. A video group costs less and can include people who live far apart. People speak more slowly online and interrupt less, and the moderator sees less of how they react. Many moderators use smaller groups online, of four to six people.
3.4 Observation
Observation means watching what people do in a setting and recording it in a systematic way. It answers a question that interviews cannot. People often do not know what they do, and they sometimes report what they wish they did.
The library students said in the groups that the room was “nice.” Observation showed that they did not sit in it.
Five Choices
A researcher who plans an observation makes five choices. Table 3.2 lists them.
| Choice | Options | The question it settles |
|---|---|---|
| Structure | Structured or unstructured | Does the observer use a checklist prepared in advance, or write open notes? |
| Disguise | Open or disguised | Do the people know they are being observed? |
| Directness | Direct or indirect | Does the observer watch the behavior, or study the traces it leaves? |
| Setting | Natural or arranged | Does the behavior take place where it normally would, or in a setting the researcher built? |
| Observer | Person or device | Does a person record the behavior, or does a camera, sensor, or software log? |
The library observation was structured, open, direct, natural, and carried out by a person. The staff member used a seating chart and marked it every 15 minutes.
Traces can be informative. Worn carpet shows where people walk. The books left on tables show what was read in the building and not checked out. Webb and colleagues (1966) called these unobtrusive measures, because collecting them does not disturb the people being studied.
A participant observer joins the activity she studies. A researcher who works a shift at a food bank learns things that a visitor with a clipboard would miss. She also becomes part of the setting, and her presence changes it.
Field Notes
The record of an observation is a set of field notes. Good notes separate three things.
- What happened. “Three students entered at 3:10. All three walked past the teen room and sat behind the reference shelves.”
- What was said, in the speaker’s words where possible.
- What the observer thought. “They seemed to be avoiding the desk.” Mark this as an interpretation.
Write the notes on the same day. Memory reduces an afternoon to a summary within hours, and the summary favors what the observer expected to see. Emerson, Fretz, and Shaw (2011) give the standard account of how to write field notes.
What Observation Cannot Show
Observation records behavior and not the reasons for it. The staff member saw where teenagers sat. She could not see why. The focus groups supplied the reason.
Observation also cannot record what happens in private, what happened in the past, or what a person intends to do.
People who know they are being watched may act differently. This is reactivity. It fades as people grow used to the observer, which is one reason to observe on many days.
3.5 Projective Techniques
Some questions fail when asked directly. A person may not know the answer. A person may know it and prefer not to say it. A projective technique asks an indirect question. The participant responds to an incomplete or ambiguous prompt, and the response shows an attitude that a direct question would miss.
Table 3.3 lists four common forms.
| Technique | What the participant does | Example |
|---|---|---|
| Word association | Says the first word that comes to mind. | “Library.” |
| Sentence completion | Finishes a sentence. | “Kids who hang out at the library are …” |
| Picture response | Describes or tells a story about a picture. | A drawing of a teenager at a library door. What is she thinking? |
| Third-person question | Answers for other people. | “Why do you think some students never go to the library?” |
The third-person question is the simplest and the most widely used. A person who will not admit to a view will often attribute it to “other people.”
A well-known study shows how the method works. Haire (1950) wanted to know why many shoppers avoided instant coffee. Asked directly, they said they disliked the taste. Haire showed two groups of women a shopping list and asked them to describe the woman who wrote it. The two lists were the same except for one item. One list had instant coffee, and the other had ground coffee. The women who saw the instant coffee list described the shopper as lazy and a poor planner far more often. The objection concerned what the product implied about the buyer, and no direct question had found it.
Reading the Responses
A researcher looks at three things across participants.
- What recurs. An ending that many people write independently is worth attention.
- How quickly people respond. A long pause can mean the person is searching for an acceptable answer.
- What people skip. A prompt that many leave blank may touch something uncomfortable.
Projective responses need more interpretation than any other data in this book. Two researchers can read the same story about a picture in different ways. Three safeguards help. Use the techniques to raise ideas and not to settle them. Have a second person read the responses, as Chapter 8 describes. Check what the responses suggest against another method.
In the library study, the sentence endings suggested that the library had a reputation as a place for students with nowhere else to go. The staff then asked about this in a later focus group. The students confirmed it.
3.6 Case Studies
A case study examines one person, organization, or event in depth, in its own setting. It is a design and not a single method. A case study of the library branch would combine the groups, the observation, staff interviews, and documents such as visit counts and floor plans.
A case study suits “how” and “why” questions about something that cannot be separated from its setting (Yin, 2018). Its findings describe the case. Whether they apply elsewhere is a judgment the reader makes from the detail the researcher provides.
A pilot study is a small trial of a method before the full study. A researcher might run one focus group to test the guide. Pilots are for fixing the method, and their data are usually not reported as findings.
3.7 Newer Forms
Several newer methods apply the same ideas in digital settings.
- Online communities. Researchers observe public forums and review sites where people already discuss a topic. Kozinets (2020) calls this netnography. Chapter 9 covers text of this kind.
- Mobile diaries. Participants record short notes, photos, or voice messages on their phones as events happen. The record is made at the time, so it depends less on memory than an interview does.
- Service maps. A researcher draws each step of a person’s dealings with a service and marks where problems occur. The map displays findings from interviews and observation. It does not replace them.
Each of these raises the questions in this chapter. Who is present and who is missing? Do people know they are being studied? What would a direct question miss?
3.8 Choosing a Method
Table 3.4 compares the four main methods.
| Depth interview | Focus group | Observation | Projective technique | |
|---|---|---|---|---|
| Best for | One person’s full account | Shared and disputed views | What people do | What people cannot or will not say |
| Private topics | Suitable | Not suitable | Depends on the setting | Suitable |
| Shows behavior | No, only reports of it | No, only reports of it | Yes | No |
| Shows reasons | Yes | Yes | No | Indirectly |
| Main risk | The interviewer leads | The group leads | The observer’s presence changes behavior | The researcher reads too much into a response |
Three questions guide the choice.
- Is the topic private? If it is, use interviews.
- Does the study need what people do, or what they say? If it needs what they do, observe.
- Can people answer a direct question? If they cannot or will not, add a projective technique.
Many studies use more than one method. The library used three, and no single one would have explained the empty room. Chapter 1 called this practice triangulation.
3.9 Work With the Data: Method Matcher
The choice of method is a judgment, and judgment improves with practice. The tool presents ten short cases. Each one describes what a researcher wants to learn.
Open Method Matcher in its own window
- Read the case. Before you choose, ask the three questions in the section above.
- Choose the method that fits best. Read the explanation, including when your choice was correct.
- Work through all ten cases. The tool reports your score at the end.
- For each case you missed, name the question that would have led you to the better answer.
Some cases could be served by two methods. When that happens, ask which method the study could least do without.
An interviewer can promise to keep what she hears confidential. A moderator cannot make that promise for a group, because eight other people also heard it. Say so at the start. Ask participants to keep the discussion private, and tell them that the researcher cannot guarantee it. Observation raises a different problem. A researcher may usually watch people in a public place without their consent. A researcher who watches in disguise, or in a place where people expect privacy, takes a serious step. That step needs a strong reason and the approval of an ethics review board.
Some services now offer “synthetic focus groups” made of invented participants that a language model plays. The output reads like a transcript. It is the model’s estimate of what such people would say, drawn from text it was trained on. No participant can surprise the moderator with an experience the model has not seen, and surprise is the reason to hold a group. A model can help draft a moderator’s guide or test whether a question is clear. It cannot stand in for the participants. A model that summarizes a real group also tends to drop who said what, and the analysis of a group depends on that.
3.10 Putting It All Together
An interview records what one person says to a researcher. Three other methods add what an interview misses. A focus group adds the reactions of other people. Observation records what people do. A projective technique asks an indirect question when a direct one would fail.
Each method has its own limit. A group shapes what its members say. Observation shows behavior and not reasons. Projective responses need careful interpretation. A study that uses two or three methods can check each one against the others.
3.11 Looking Ahead
The method is chosen. The next decision is whom to include. A survey draws a random sample. A qualitative study chooses its participants on purpose, and the reasons for each choice matter. Chapter 4 covers how to choose and how to know when enough people have been heard.
3.12 Further Reading and References
- Krueger, R. A., and Casey, M. A. (2015). Focus Groups: A Practical Guide for Applied Research (5th ed.). Sage. The standard practical guide.
- Morgan, D. L. (1997). Focus Groups as Qualitative Research (2nd ed.). Sage.
- Guest, G., Namey, E., and McKenna, K. (2017). How many focus groups are enough? Building an evidence base for nonprobability sample sizes. Field Methods, 29(1), 3 to 22.
- Emerson, R. M., Fretz, R. I., and Shaw, L. L. (2011). Writing Ethnographic Fieldnotes (2nd ed.). University of Chicago Press.
- Webb, E. J., Campbell, D. T., Schwartz, R. D., and Sechrest, L. (1966). Unobtrusive Measures: Nonreactive Research in the Social Sciences. Rand McNally.
- Haire, M. (1950). Projective techniques in marketing research. Journal of Marketing, 14(5), 649 to 656.
- Yin, R. K. (2018). Case Study Research and Applications: Design and Methods (6th ed.). Sage.
- Kozinets, R. V. (2020). Netnography: The Essential Guide to Qualitative Social Media Research (3rd ed.). Sage.
3.13 Key Terms
- Case study. A study of one person, organization, or event in depth, in its own setting.
- Conformity. Agreeing with a group and keeping one’s own view private.
- Depth interview. A one-to-one interview that explores a topic in detail.
- Field notes. The written record of an observation.
- Focus group. A guided discussion among six to ten participants.
- Moderator. The person who leads a focus group.
- Netnography. The study of online communities through what their members post.
- Observation. Watching what people do and recording it in a systematic way.
- Participant observer. A researcher who joins the activity she studies.
- Pilot study. A small trial of a method before the full study.
- Projective technique. An indirect question that asks a participant to respond to an incomplete or ambiguous prompt.
- Reactivity. A change in behavior that occurs because people know they are being watched.
- Segment. A set of participants who are alike on a characteristic that matters to the study.
- Unobtrusive measure. Evidence collected with no disturbance to the people studied, such as a trace they left behind.
3.14 Exercises
Check Your Understanding
- What does a focus group show that eight separate interviews would not?
- Why is one focus group never enough?
- Name the five choices in the design of an observation.
- Why did the shoppers in the coffee study give one reason when asked directly and reveal another when asked to describe a shopper?
Apply It
- Write a moderator’s guide with the four parts in Table 3.1 for a focus group of bus riders on a route that may be cut.
- Observe a public place, such as a cafe or a bus stop, for 20 minutes. Write field notes that separate what happened, what was said, and what you thought.
- Write three sentence-completion prompts for a study of why employees do not use a company’s free counseling service.
- A clinic wants to learn why patients miss appointments. Choose two methods from this chapter and state what each would show that the other would not.
Think Deeper
- A manager watches four focus groups from behind a one-way mirror. The manager then reports that customers love the new design. What should the researcher ask before accepting that statement?
- A researcher wants to observe how staff treat customers and plans to pose as a customer. What is gained, and what is the ethical cost? Under what conditions, if any, would you approve the study?