1  What Words Can Show

Where this chapter fits. Steps 1 and 3 of the research process are to formulate the problem and to choose the design. The Introduction describes all eight steps. This chapter explains which questions call for qualitative research and what such a study can claim.

1.1 The Checkbox That Said “Other”

Consider a school district that has lost 6 percent of its students in three years. The case is a composite written for this book, and the district is not a real one. Its details are typical of what district staff describe.

The district has good numbers. It knows how many students left each school, in which grade, and in which month. When a family withdraws a child, the office staff fills in a form. The form has a box labeled “reason for leaving,” with five choices. Sixty percent of the forms are marked “moved out of district.” Thirty percent are marked “other.”

The school board asks the superintendent why families are leaving. The numbers do not answer the question. “Moved” does not say why the family moved. “Other” says nothing at all.

A staff member proposes a different approach. She will interview 15 families who left in the past year. She will ask each one to describe the decision in their own words. Three excerpts from those interviews follow. They are synthetic.

Synthetic “We didn’t want to leave. They cut the bus route. I start work at six. There was no way to get him there.”

Synthetic “The rent went up 300 dollars. We found a place we could afford, and it was across the district line. That’s the whole reason. We loved that school.”

Synthetic “She has a reading plan. The teacher who ran it left, and nobody picked it up. I asked four times. The other district had a spot.”

Each excerpt names a cause that the form could not record. One is transportation. One is the cost of housing. One is a service the school stopped providing. The first and third are things the district can act on. The second is outside its control.

The 15 interviews cannot show what share of all departing families left for each reason. They can show what the reasons are and how families weighed them. They can also show which reasons the district’s own form was hiding. This chapter explains what kind of knowledge that is and when a researcher should seek it.

1.2 What Qualitative Research Is

Qualitative research studies how people experience and interpret some part of their lives. Its data are words, and sometimes images and observed behavior. The researcher collects those data through interviews, group discussions, observation, and documents. The researcher then analyzes them by reading closely, labeling what is there, and looking for patterns.

Quantitative research records characteristics as numbers and analyzes them with statistics. Each approach fits some questions better than others. Table 1.1 sets them side by side.

Table 1.1: Quantitative and qualitative research compared.
Quantitative research Qualitative research
Typical question How many? How much? Is there a difference? How? Why? What is it like?
Data Numbers from surveys, records, and experiments Words from interviews, discussions, observation, and documents
Participants Many, often chosen at random Few, chosen on purpose for what they can explain
Instrument A fixed questionnaire A flexible guide, with the researcher asking follow-up questions
Analysis Statistics Coding and the search for themes
Typical finding An estimate with a margin of error A pattern, with quotations as evidence
Main risk Measuring the wrong thing with precision Seeing a pattern the data do not support

Neither approach is the more rigorous one. Each has standards, and each can be done badly. A survey with a biased sample is poor research. So is a set of interviews in which the researcher heard only what she expected to hear.

1.3 Questions That Call for Words

Qualitative research fits four kinds of questions.

Questions about process. How does a family decide to change schools? What steps does a tenant go through after receiving a notice? A survey can record the outcome. An interview can trace the sequence.

Questions about meaning. What does “stable housing” mean to a person who has moved five times? Two people can give the same answer on a survey and mean different things by it.

Questions where the categories are unknown. The district’s form offered five reasons for leaving. The interviews produced reasons that were not among the five. When a researcher does not yet know the possible answers, a fixed list of choices will miss some of them.

Questions about people who are hard to count. Some groups do not appear in records or do not respond to surveys. A researcher can sometimes reach them through trusted contacts and a conversation.

Qualitative research is a poor fit for questions of how many and how much. Fifteen interviews cannot show that 40 percent of families left because of bus routes. A report that makes such a claim from 15 interviews has gone beyond its data.

1.4 A Well-Known Example

In 2008 and 2009, the sociologist Matthew Desmond lived in two poor neighborhoods of Milwaukee. He lived first in a trailer park and then in a rooming house. He followed eight families and two landlords through evictions, moves, and court hearings. His book Evicted (2016) reports what he observed.

The fieldwork showed him something that court records did not. Many families were forced out with no court case at all. A landlord would tell a tenant to leave, or pay the tenant a small sum to go, or remove the front door. None of these events appeared in any record.

Desmond then designed a survey of about 1,100 Milwaukee renters to count what he had seen. The survey found that informal evictions were about twice as common as formal ones. Court records had been missing most forced moves.

The order of the two studies matters. The fieldwork found a category that no dataset contained. The survey then measured how large it was. Without the fieldwork, the survey would not have included the question.

1.5 What a Qualitative Study Can Claim

A probability sample supports a claim about a population. A qualitative study does not, and it should not try to. Its claims are of other kinds.

A claim that something exists. The district’s interviews establish that some families left because a bus route was cut. One clear case is enough to show that a reason exists.

A claim about how something works. Interviews can show the steps that connect a cause to an outcome. A rent increase leads to a search. The search finds a cheaper home across the district line. The child then changes schools.

A claim about range. A well-chosen set of participants can show the variety of experiences in a group, including ones the researcher did not expect.

A claim that may transfer. A reader in another district can judge whether the findings are likely to apply there. The researcher supports that judgment by describing the setting and the participants in detail. Lincoln and Guba (1985) call this property transferability. It is the reader, and not the researcher, who decides whether the findings transfer.

A qualitative report should use words that match these claims. “Several families described” and “one pattern in the interviews was” are accurate. “Most families feel” is a claim about a population, and the study cannot support it.

1.6 Five Traditions

Qualitative researchers work in several traditions. Creswell and Poth (2018) describe five. Table 1.2 summarizes them.

Table 1.2: Five traditions of qualitative research, after Creswell and Poth (2018).
Tradition What it studies Typical data
Narrative research The life story of one person or a few people Long interviews, letters, diaries
Phenomenology What an experience is like for those who have lived it In-depth interviews with people who share the experience
Grounded theory A process, with the aim of building a theory from the data Interviews collected and analyzed in rounds
Ethnography The shared practices of a group or a place Long-term observation, with interviews
Case study One bounded case, such as a program or an organization Interviews, documents, and observation together

Much applied research in business, health, and public policy does not follow one tradition closely. It uses interviews or focus groups to answer a practical question and analyzes them for themes. This book teaches that general approach. The methods in the chapters ahead are common to all five traditions.

1.7 The Four Sources of Data

Qualitative data come from four main sources.

An interview is a guided conversation between a researcher and one participant. Chapter 2 covers it in detail. A focus group is a guided discussion among six to ten participants. It shows how people respond to one another’s views. Observation means watching what people do in a setting and recording it in field notes. Documents include anything already written, such as letters, case files, public comments, and posts online. Chapter 8 covers documents.

Each source has a blind spot. Interviews record what people say they do, which can differ from what they do. Observation records behavior and not the reasons behind it. Documents were written for a purpose other than research. Researchers often combine sources so that each one can check the others. This practice is called triangulation, and Chapter 7 returns to it.

1.8 Work With the Data: Read Two Interviews

This book uses one set of 12 interviews about housing in every chapter. The interviews are synthetic. They were written for this book, and no real person said these words. Read two of them now, before Chapter 2.

The interviews are plain text files. Start with H01 and H06.

  1. Read each interview once from start to finish.
  2. In each one, mark a statement that a survey question would have been unlikely to record.
  3. Write one question you would want to ask the participant next.

Keep your notes. Chapter 5 returns to these interviews for coding.

1.9 Using Words and Numbers Together

A study that combines qualitative and quantitative methods is a mixed methods study. Creswell and Plano Clark (2018) describe three basic designs.

  1. Words first, then numbers. Interviews identify the categories, and a survey measures how common each one is. Desmond’s work followed this order. So would the school district, if it rewrote its withdrawal form after the interviews.
  2. Numbers first, then words. A survey finds a pattern, and interviews explain it. A district that sees withdrawals rise sharply in one school could interview families from that school.
  3. Both at once. The two kinds of data are collected in the same period and compared. If they agree, confidence in the finding grows. If they disagree, the researcher has a new question.

The choice depends on what is already known. If the categories are unknown, begin with words. If a pattern needs to be explained, begin with numbers.

1.10 The Researcher Is Part of the Study

In a survey, every respondent receives the same questions in the same order. In an interview, the researcher decides what to ask next. In the analysis, the researcher decides what a passage means. The researcher’s judgment is part of the method.

This has a cost. Two researchers can read the same transcript and notice different things. A researcher’s background, beliefs, and hopes for the study all affect what she hears.

Qualitative research manages this cost in the open. Reflexivity is the practice of examining how the researcher’s own position affects the study, and of reporting it. A former teacher who interviews parents about schools should say so in the report. She should also consider which answers she may have been ready to hear.

The other safeguards are procedures. They include written rules for labeling the data, a second person who labels the same material, and a record of each decision. They also include a search for cases that contradict the pattern. Chapters 5 through 7 cover each of these.

Ethics Moment: Whose Account Is It

An interview gives a researcher someone else’s words. The researcher then selects, cuts, and arranges them. A participant who said ten things may find one of them in the report, placed next to a claim the participant would not accept. The researcher has a duty to represent each person’s account fairly. Keep quotations in context. Do not join parts of two answers into one quotation. When a participant’s view does not fit the pattern, report it.

AI Reality Check: Invented Quotations Are Not Data

A language model can write an interview transcript on any topic in seconds. The transcripts in this book were drafted that way, and each one is marked as synthetic. They are teaching material. They are not evidence about housing, and no finding about real people can rest on them. The same limit applies to any text a model produces. If you ask a model what parents think about school bus routes, it will write fluent and plausible answers. No parent said them. Qualitative evidence begins with a real person who said or wrote something. A model cannot supply that step.

1.11 Putting It All Together

Qualitative research studies how people experience and interpret their lives. Its data are words from interviews, focus groups, observation, and documents. It fits questions about process and meaning. It also fits questions where the possible answers are not yet known.

A qualitative study does not estimate how common something is. It can show that something exists, how it works, and what range of experiences a group contains. A reader judges whether the findings transfer to another setting.

Words and numbers can be combined. Interviews can supply the categories that a survey then counts, and interviews can explain a pattern that a survey found. The researcher’s judgment is part of the method, and reflexivity and written procedures keep that judgment open to inspection.

1.12 Looking Ahead

The staff member in the school district now has to conduct 15 interviews. What she asks, and how she asks it, will shape everything that follows. Chapter 2 covers the interview. The case is a set of conversations with parents who used a crisis nursery.

1.13 Further Reading and References

  • Creswell, J. W., and Poth, C. N. (2018). Qualitative Inquiry and Research Design: Choosing Among Five Approaches (4th ed.). Sage. The source of the five traditions.
  • Creswell, J. W., and Plano Clark, V. L. (2018). Designing and Conducting Mixed Methods Research (3rd ed.). Sage.
  • Desmond, M. (2016). Evicted: Poverty and Profit in the American City. Crown. The appendix, “About This Project,” describes the fieldwork and the survey.
  • Desmond, M., and Shollenberger, T. (2015). Forced displacement from rental housing: Prevalence and neighborhood consequences. Demography, 52(5), 1751 to 1772. The survey results on formal and informal evictions.
  • Lincoln, Y. S., and Guba, E. G. (1985). Naturalistic Inquiry. Sage. The source of the term transferability.
  • Tracy, S. J. (2020). Qualitative Research Methods (2nd ed.). Wiley-Blackwell. A readable general textbook.

1.14 Key Terms

  • Documents. Text that already exists, such as letters, case files, and public comments.
  • Field notes. A researcher’s written record of what was observed.
  • Focus group. A guided discussion among six to ten participants.
  • Interview. A guided conversation between a researcher and one participant.
  • Mixed methods. A study that combines qualitative and quantitative methods.
  • Observation. Watching what people do in a setting and recording it.
  • Qualitative research. The study of how people experience and interpret some part of their lives, using words as data.
  • Quantitative research. Research that records characteristics as numbers and analyzes them with statistics.
  • Reflexivity. Examining and reporting how the researcher’s own position affects the study.
  • Transferability. The degree to which findings may apply in another setting, as judged by the reader.
  • Triangulation. The use of more than one source or method to check a finding.

1.15 Exercises

Check Your Understanding

  1. Give two kinds of questions that qualitative research fits well and one kind that it fits poorly.
  2. The district interviewed 15 families. Which of these claims can the interviews support? (a) Some families left because a bus route was cut. (b) Bus routes were the most common reason for leaving. (c) Families weighed the commute against the quality of a program.
  3. What did Desmond’s fieldwork find that court records had missed? Why did the survey come after the fieldwork?
  4. Define reflexivity. Give one example of a researcher’s background that should be reported.

Apply It

  1. Rewrite the district’s withdrawal form. Use the three synthetic excerpts to add at least three new choices for “reason for leaving.”
  2. For each question, name the mixed methods design that fits best and say why. (a) A clinic’s survey shows that patients over 70 miss more appointments. (b) A city wants to survey residents about a new park and does not know which features they care about.
  3. Read transcript H05 or H09 in the corpus. Write two sentences that report what the participant said, using wording that matches what one interview can support.

Think Deeper

  1. A report based on 12 interviews states, “Most renters feel that landlords do not listen.” Rewrite the sentence so that it claims only what the study can support. What was lost, and what was gained?
  2. A researcher who once lost her own home plans to interview people facing eviction. How could her experience help the study? How could it distort it? What should she do about both?
  3. Choose a number from your own work or studies that you do not fully understand. Describe a small qualitative study that could help explain it. Whom would you talk to, and what would you ask?