How do you study what numbers cannot measure?

Start here

What can a questionnaire not tell you?

A questionnaire counts how many people did something. It cannot tell you what the experience meant to them, or why they acted as they did, unless you already knew the possible answers when you wrote it.

Introducing a BMJ series in 1995, Pope and Mays wrote that qualitative and quantitative research tend to be portrayed as opposites, and set out to show how qualitative methods can complement quantitative research.

Comparison

Qualitative or quantitative: how do they differ, and when do you choose each?

Quantitative research asks how many, how much and how often. Qualitative research asks what, how and why.

Choose qualitative methods when

• Little is known about the topic and you do not yet know which questions to ask.
• You need people's experiences, beliefs or reasons in their own words.
• You need to understand how a programme or process actually works in its setting.
• You want to word a questionnaire, or explain a survey result.

Choose quantitative methods when

• You need a number for a population: a prevalence, an average or a risk.
• You want to test a hypothesis or compare treatments.
• You need results that apply beyond the people studied, with a known margin of error.

Many studies combine the two (mixed methods). A focus group discussion of a questionnaire is useful for testing how its questions are worded and for explaining survey results (Kitzinger, 1995). The ICMR 2017 guidelines note that implementation research often uses mixed quantitative and qualitative methods. See the study designs chapter for quantitative designs.

Interactive

Sort research questions into qualitative, quantitative or mixed methods

How many, how much and how often need numbers. What, how and why need words. A question that needs both needs mixed methods.

Designs

What are the main qualitative designs?

Ethnography, phenomenology and grounded theory are often called the 'big three' qualitative approaches. The case study is a fourth design, used to study a programme, policy or event in its real-life setting. The research question decides which one fits.

Ethnography

The study of the culture of a group: its shared behaviour, beliefs and practices. The researcher spends long periods in the setting, observes, holds informal conversations and interviews key informants, and writes a detailed description of the culture (Korstjens and Moser, 2017). Bronisław Malinowski's Argonauts of the Western Pacific (1922) was based on months of living among the Trobriand Islanders, watching their daily work and speaking with them in their own language.

Phenomenology

The study of lived experience: what it is like to have a condition or go through an event, from the person's own point of view. Participants are chosen because they have had the experience. Data come mainly from in-depth interviews, and the findings are themes that describe the essential meaning of the experience (Moser and Korstjens, 2017). As a philosophy, phenomenology was launched by Edmund Husserl in Logical Investigations (1900–01). In later work he proposed “bracketing” the question of whether the outside world exists, in order to study the structure of experience itself.

Grounded theory

Builds a theory from the data instead of testing one chosen in advance. Barney Glaser and Anselm Strauss described it in The Discovery of Grounded Theory (1967). Collection and analysis run together. Each new piece of data is compared with what has already been coded (the constant comparative method), and the next participants are chosen for what they can add to the developing theory (theoretical sampling). Sampling stops at theoretical saturation, which Glaser and Strauss defined as the point where “no additional data are being found whereby the sociologist can develop properties of the category”.

Case study

An in-depth study of one case, or a few, in its real-life context: a patient's episode of care, a ward, or a hospital's new service. The case has a defined boundary in time and place, and several data sources are combined (Crowe et al., 2011). Stake described three types: intrinsic (a case studied because it is unique), instrumental (a case used to understand a broader issue) and collective (several cases studied together).

Other approaches include narrative research, conversation analysis, historical research, participatory action research, and descriptive studies analysed by thematic or content analysis.

Decide

Which design fits your question?

Answer one or two questions. Each answer links to the section that explains it.

Data collection

How do you collect qualitative data?

The three most used methods are participant observation, in-depth interviews and focus group discussions. Many studies combine more than one.

Focus group discussions

How do you run a focus group discussion?

A focus group is not a group interview in which each person answers in turn. The method uses the discussion between participants as data.

• Size: most focus groups have 6 to 12 participants. Smaller groups of 4 to 6 suit topics where each person has an intensive experience to share (Wong, 2008).
• Composition: people who share a key characteristic talk more freely. Keep groups separate when one group would stop another from speaking, for example younger and older women discussing sexual behaviour (Wong, 2008).
• Staff: a moderator guides the discussion. A note-taker records what was said, who spoke in what order, and the tone of the discussion.
• Questioning route: open-ended questions, general ones first and specific ones later, tested in a mock group (Moser and Korstjens, 2017).
• Duration: 45 to 90 minutes in one guide (Wong, 2008), 90 to 120 minutes in another (Moser and Korstjens, 2017).
• Number of groups: continue until no new themes appear. In empirical tests, saturation came within 4 to 8 focus groups (Hennink and Kaiser, 2022).

What the moderator does

1. Explains the purpose, the ground rules, how the session is recorded, and who will see the information.
2. Asks participants to speak one at a time, and says there are no right or wrong answers.
3. Uses probes such as “Can you tell me more?” or “Would you give me an example?”.
4. Watches the group so that everyone takes part. When some people dominate, addresses questions to those who have not spoken.
5. Brings the discussion back when it moves away from the topic, and keeps to time.
6. Stays neutral, because the moderator's own views can sway the group.

Teaching scenario: in a group of eight mothers discussing missed measles vaccination, one mother says loudly that the vaccine causes fever, and the others go quiet. The moderator thanks her, asks “Has anyone had a different experience?”, and turns to a mother who has not yet spoken.

In-depth interviews

How do you conduct an in-depth interview?

A one-to-one conversation that explores one person's experiences and views on a set of issues in detail.

• Types: structured (fixed questions in fixed order), semi-structured (a guide of topics and open questions) and unstructured (a few broad topics). Most qualitative interviews are semi-structured.
• Interview guide: open questions, then follow-up questions. Probes are exploratory questions such as “Then what happened?”. Prompts are signs that encourage the person to go on, such as eye contact and leaning forward. Pilot the guide first.
• Length: 30 to 90 minutes, audio-recorded with consent and transcribed word for word.
• Key informant interviews: interviews with people who have special knowledge of the topic and can help you gain access to others.
• Problems to plan for: gatekeepers, finding a private place, socially desirable answers, and participants reluctant to talk (Moser and Korstjens, 2017).

Observation

What does observation mean in qualitative research?

It is not clinical observation of a patient. The researcher systematically watches people and events in their natural setting to understand behaviour and interactions, and writes field notes (Mays and Pope, 1995).

Sampling

Whom do you study, and how many?

Participants are chosen on purpose, for the information they can give. The number is decided by saturation, not by a power calculation.

The usual strategies are purposive, criterion, theoretical, snowball and convenience sampling. The sampling chapter describes each method. Ethnography usually samples key informants, phenomenology samples people who have had the experience, and grounded theory moves from purposive to theoretical sampling (Moser and Korstjens, 2017).

Data saturation is the point at which new data give no new analytical information. It is common to carry out two or three more interviews to confirm it (Moser and Korstjens, 2017). Published tests give a guide to numbers:

• Guest and colleagues (2006), analysing 60 interviews, found saturation within the first 12.
• Hennink and colleagues found that 9 interviews identified the range of issues (code saturation), but 16 to 24 were needed to understand them in depth (meaning saturation).
• A systematic review found saturation within 9 to 17 interviews or 4 to 8 focus groups, mostly in studies of fairly similar participants with narrow objectives (Hennink and Kaiser, 2022).

In the protocol, state the sampling criteria, an estimated number with its basis, and how you will judge saturation.

Interactive

Add interviews one at a time and watch the number of new codes fall

Each interview is compared with the codes already found. Data collection stops when further interviews add nothing new.

Analysis

What do you do with hours of transcripts?

Analysis starts after the first interview, not after the last. You label parts of the text with codes, group codes into categories, and build themes that answer the research question.

The six phases of thematic analysis (Braun and Clarke, 2006)

1. Familiarising yourself with your data: transcribe, read and re-read, and note initial ideas.
2. Generating initial codes: code interesting features systematically across the whole data set.
3. Searching for themes: collate codes into possible themes.
4. Reviewing themes: check each theme against the coded extracts and the whole data set.
5. Defining and naming themes: refine what each theme covers and give it a clear name.
6. Producing the report: choose clear extracts and relate the analysis to the research question and the literature.

The phases are not strictly in sequence. You move back and forth between them. Grounded theory uses constant comparison with open, axial and selective coding. The framework approach and content analysis are other methods.

Software

Software stores, codes, searches and retrieves text, but you still decide the codes and themes (Moser and Korstjens, 2017). Programs include NVivo, ATLAS.ti and MAXQDA. QDA Miner Lite is free, and Taguette is free and open source.

Interactive

Assign interview quotes to themes and see the codes group together

Each quote gets a code for its idea. Codes from different patients that share a pattern of meaning form a theme.

Quality

How do you show that qualitative findings can be trusted?

Lincoln and Guba (1985) proposed four criteria for trustworthiness, as equivalents of the terms used in quantitative research.

Reflexivity means stating your own background and position and considering how they influenced data collection and interpretation (Korstjens and Moser, 2017). Mays and Pope (1995) also describe coding transcripts by more than one researcher and feeding findings back to participants.

Triangulation

What is triangulation?

Studying the same question with more than one method, data source, investigator or theory, to build a fuller understanding and to check whether the findings agree.

Denzin and Patton identified four types (Carter et al., 2014):

• Data source triangulation: different people, times or places, for example patients, relatives and health workers.
• Investigator triangulation: more than one researcher collects or analyses the data.
• Theory triangulation: more than one theory is used to interpret the data.
• Method triangulation: more than one method, for example interviews, focus groups and record review.

Lincoln and Guba list triangulation as one way to establish credibility. In case studies, data collected in different ways are expected to lead to similar conclusions (Crowe et al., 2011).

Social network analysis

How do you study relationships between people or organisations?

Social network analysis maps who is connected to whom. People or organisations are nodes, and the relationships between them are ties.

Hawe and Ghali (2007) mapped staff relationships in a school. The person named by the most colleagues (highest degree centrality) was not always the best choice of intervention champion. To reach the most marginal people, betweenness centrality could be the better guide.

The centrality and density scores are numbers. In qualitative social network analysis, participants draw or describe their networks in interviews and explain the relationships. In a South African study, hospital staff used Net-Map, marking actors, arrows for relationships and towers for influence, to show who drove the implementation of national quality standards (Mabena and Armstrong, 2025).

Strengths and limits

What can qualitative studies do, and what can they not do?

Strengths

• Explain the reasons, meanings and processes behind numbers.
• Study topics where little is known, and generate hypotheses and questionnaire items.
• Take account of the setting, and remain open to findings nobody expected.
• Include people who cannot read or write, and people reluctant to be interviewed alone (focus groups).
• Allow the design to change as findings emerge.
• Need small samples, so they are feasible in one unit or community.

Limits

• Cannot give a prevalence or an effect size, and findings cannot be generalised statistically.
• The researcher's views can influence the data and the interpretation.
• Data collection, transcription and analysis take a long time.
• Quality depends on the skill of the interviewer or moderator.
• Participants may give socially desirable answers, or behave differently when observed.
• Full confidentiality is harder, because quotes and small settings can identify people.

By specialty

Qualitative questions by specialty

Common myths

Common mistaken beliefs

• ✗Myth: Qualitative research is only opinion.

  ✓It has defined designs, sampling strategies, analysis methods and quality criteria, and reporting checklists (COREQ, SRQR).

• ✗Myth: A study of 15 interviews is invalid because the sample is small.

  ✓The aim is depth, not a prevalence. Sample size is judged by saturation, which empirical tests reached within 9 to 17 interviews.

• ✗Myth: Qualitative findings apply to the whole population.

  ✓They cannot be generalised statistically. Readers judge whether they transfer to their own setting, using the description of participants and context.

• ✗Myth: A focus group is a quick way to interview many people at once.

  ✓A focus group uses the discussion between participants as data. Asking each person in turn is a group interview.

• ✗Myth: Software does the analysis.

  ✓Software stores and retrieves coded text. The researcher decides the codes, categories and themes.

• ✗Myth: Triangulation means using exactly three methods.

  ✓It means more than one method, data source, investigator or theory. The name comes from surveying, not from the number three.

Why this matters to you

Five consequences of getting qualitative methods wrong

1. Protocol sent back

   The committee asks how you will choose participants and why that number. State purposive criteria and how you will judge saturation.

2. A results chapter that is a list of quotes

   Quotes arranged under interview questions, with no codes or themes, do not show how you reached your conclusions.

3. Questions you cannot answer at the viva

   What is saturation? How is a focus group different from a group interview? How did you establish credibility?

4. Misreading papers

   Judging an interview study by its p-values, or accepting one that does not say who collected the data and how themes were derived.

5. Programmes that miss the reasons

   A service planned from numbers alone may not address the reasons patients give for not using it.

History

History of qualitative research methods

1. 1900–01

   Edmund Husserl

   Logical Investigations launched phenomenology as a philosophy.

2. 1922

   Bronisław Malinowski

   Argonauts of the Western Pacific, based on months of fieldwork among the Trobriand Islanders, set out his methods of fieldwork.

3. 1946

   Robert Merton and Patricia Kendall

   Described "the focused interview", designed to find out how people responded to a situation the investigator had already analysed.

4. 1967

   Barney Glaser and Anselm Strauss

   The Discovery of Grounded Theory introduced theory built from data, the constant comparative method and theoretical saturation.

5. 1978

   Norman Denzin

   The 1978 edition of The Research Act (first published 1970) is cited for four types of triangulation.

6. 1985

   Yvonna Lincoln and Egon Guba

   Naturalistic Inquiry proposed credibility, transferability, dependability and confirmability as equivalents of internal validity, external validity, reliability and objectivity.

7. 1995

   Catherine Pope and Nicholas Mays, BMJ

   Began a BMJ series showing how qualitative methods can complement quantitative research, with "Reaching the parts other methods cannot reach".

8. 2006

   Virginia Braun and Victoria Clarke

   Set out six phases of thematic analysis.

9. 2007

   Allison Tong, Peter Sainsbury and Jonathan Craig

   Published COREQ, a 32-item checklist for reporting interview and focus group studies.

10. 2022

    Monique Hennink and Bonnie Kaiser

    A systematic review of 23 articles found saturation within 9 to 17 interviews or 4 to 8 focus groups in empirical tests.

Guidelines

Qualitative research in guidelines

Steps

Doing an interview or focus group study, step by step

1. Write a broad, open research question and choose the design that fits it.
2. Define whom you need and why: the setting, the purposive criteria, and an estimated number.
3. Prepare the interview guide or questioning route, and pilot it.
4. Obtain ethics approval. Plan consent, audio-recording, confidentiality and what to do if a participant becomes distressed.
5. Collect data, record it, transcribe word for word, and write field notes.
6. Start analysis after the first interview: code, group into categories, build themes, and write memos.
7. Continue until new data add no new codes, then confirm with two or three more.
8. Check trustworthiness: triangulation, member checking, an audit trail and a reflexive account.
9. Report the study using COREQ or SRQR.

Reading a paper

Five checks on any qualitative study

• Does the research question need a qualitative approach?
• Were participants chosen on purpose, with stated criteria, and is saturation explained?
• Do the authors say who collected the data, with what guide, and how it was recorded and transcribed?
• Is it clear how codes and themes were derived, and do the quotes support the themes?
• Did the authors address reflexivity, use triangulation or member checking, and follow COREQ or SRQR?

Why is it called that?

Word surgery

Exam answers

Model answer outlines

Past papers

Every TNMGRMU question on qualitative research

Each one is answered in this chapter. Tap a sitting to open the university's own question paper.

1. ×2Focus groupsPh.D. RM · Jan 2018Ph.D. RM · Jan 2016
2. ×1What are the Qualitative Research methods? Enumerate the role in Public healthMD Community Medicine · May 2025
3. ×1Discuss the role of qualitative research methods in public healthMD Community Medicine · Oct 2025
4. ×1Classify Quantitative Research Designs. Discuss the Differences between qualitative and qualitative research approaches. Why is design important in Research studies?Ph.D. RM · Jul 2025
5. ×1Ethnographic researchPh.D. RM · Jul 2025
6. ×1Outline the strength and weakness of Focus Group Discussion (FGD). Discuss the qualities of facilitator of FDG with suitable scenarioMD Community Medicine · May 2024
7. ×1Discuss the differences between qualitative and quantitative research approaches. When and why might one be preferred over the other?Ph.D. RM · Jul 2024
8. ×1Concept of triangulation in ResearchPh.D. RM · Jul 2024
9. ×1Describe “Social Network Analysis” as a Qualitative Research Tool for Public Health interventionsMD Community Medicine · May 2023
10. ×1Describe the study designs for Qualitative research? What are the tools that can be used for qualitative studies? List the various software available for Analysing Qualitative DataMD Community Medicine · Oct 2023
11. ×1In depth interviewsPh.D. RM · Jan 2023
12. ×1FGD (Focus Group Discussion)Ph.D. RM · Jul 2023
13. ×1Quantitative vs Qualitative researchMD Community Medicine · Oct 2022
14. ×1Focus Group discussionsPh.D. RM · Aug 2021
15. ×1Advantages of qualitative studyPh.D. RM · Jul 2019
16. ×1EthnographyPh.D. RM · Jul 2016
17. ×1Qualitative ResearchMD Community Medicine · Apr 2015
18. ×1Ethnography designPh.D. RM · Jan 2015
19. ×1PhenomenologyPh.D. RM · Jul 2015
20. ×1List the various types of epidemiological study designs. Write role of qualitative researchMD Community Medicine · Apr 2014
21. ×1Discuss the types of qualitative research designPh.D. RM · Jul 2014