Choosing a research method: qualitative or quantitative?
Your research question decides the method, not the other way round: questions starting with how many or to what extent need a quantitative design, questions starting with how do people experience or why need a qualitative one. Quantitative work measures a few variables in many cases and can generalise; qualitative work examines many aspects of a few cases and can explain. Mixed methods is not a compromise but a design in its own right, and it requires you to justify why both are needed.
Key takeaways
- The research question determines the method – finish the question before you choose.
- Quantitative gives breadth and generalisability, qualitative gives depth and explanation.
- Sampling is not a detail: it decides what your results are allowed to claim.
- The methods chapter must be precise enough for someone else to repeat the study.
What is a methods chapter for?
The methods chapter is the systematic account of how you collected and analysed your material. It serves two purposes: it makes the study auditable, and it makes it repeatable. A methods chapter that fails either test is too vague, however well written it is.
Method is therefore not the same as methodology. Methodology is the theoretical justification for the choice; method is what you actually did, in what order, with whom, and with which instrument.
How do I know if my question is quantitative or qualitative?
Look at the verb in the question. Questions containing measure, compare, association, proportion or effect are quantitative. Questions containing experience, describe, interpret, understand or meaning are qualitative.
If the test does not resolve it, the question is usually too broad. "How does online teaching affect students?" cannot be answered by any method at all. "How do nursing students experience remote supervision in their fourth term?" is qualitative and answerable. "What proportion of nursing students at one university completed the course online compared with on campus?" is quantitative and answerable.
What separates the two in practice?
| Quantitative | Qualitative | |
|---|---|---|
| Purpose | Measure, compare, generalise | Understand, interpret, explain |
| Data collection | Survey, registry, experiment, measurement | Interview, observation, text analysis |
| Sampling | Random or representative, usually large | Purposive, usually 5–20 |
| Analysis | Descriptive and inferential statistics | Thematic analysis, content analysis, grounded theory |
| Quality concepts | Validity and reliability | Credibility, transferability, dependability |
| Reach of findings | The population the sample represents | The phenomenon in its context |
How big does my sample need to be?
For a qualitative interview study at undergraduate level the normal range is five to fifteen participants, and what settles it is saturation – the point where new interviews stop producing new themes. Guest, Bunce and Johnson (2006) found in a widely cited study that most themes had already emerged within the first twelve interviews in a homogeneous group.
For a quantitative study, size is determined by the effect you want to be able to detect, not by what feels reasonable. Run a power calculation in advance and state in the methods chapter which effect size the study was powered for. An undersized sample does not give you a cautious result; it gives you an uninterpretable one.
Which method suits which question?
Interviews suit questions about reasoning and experience, and semi-structured is the undergraduate default because it keeps answers comparable without closing them down. Surveys suit closed questions where you need breadth, but the wording has to be piloted on a handful of people first. Observation suits questions about behaviour rather than perception, because people systematically describe their behaviour differently from how they perform it. Document and text analysis suits questions about how something is represented in material that already exists.
How do I analyse qualitative data?
Thematic analysis as set out by Braun and Clarke (2006) is the usual route, in six phases: familiarise yourself with the data, generate codes, search for themes, review themes, define them, and write them up. The crucial difference from a summary is that themes must answer the research question rather than mirror the headings of your interview guide. If your themes carry the same names as your interview questions, you have sorted the material, not analysed it.
What goes into the methods chapter?
Six parts, in this order: design, sampling and inclusion criteria, data collection with dates and instruments, analytical approach, ethical considerations, and a methods discussion. The ethics section is not a formality – informed consent, anonymisation and data handling must be stated concretely. Write in the past tense and in the voice your department expects; the passive is still standard in health sciences, while the active voice is gaining ground in the social sciences.
What are the most common mistakes?
Four recur. The first is choosing the method before the question, which shows up as a research question quietly rewritten to suit a survey. The second is a convenience sample not reported as a convenience sample. The third is qualitative findings presented as percentages, which creates false precision – write "most participants" rather than "67 per cent" when n is twelve. The fourth is a methods discussion that lists only strengths; a credible discussion names at least one weakness and says what it means for the conclusion.
How do I get started?
Write the research question as a single sentence and delete everything in it that cannot be investigated. Then find three high-marked dissertations in your subject and read only their methods chapters – they show what your department actually expects, faster than any methods textbook. Check that your material is obtainable before you lock the design, and read on in our dissertation guide and our guide to evaluating sources.
Methods literature is heavy going. Upload your course's methods textbook to Memmo and let the chat explain the difference between two concepts using your own material as the starting point – Memmo is used by more than 75,000 students across over 5,000 courses.
Frequently asked questions
Can I combine qualitative and quantitative methods?
Yes, that is mixed methods, but it requires you to justify why both are needed and to describe how they relate – sequentially or in parallel. At undergraduate level the workload is routinely underestimated; two half-finished sub-studies mark lower than one done well.
How many interviews are enough for an undergraduate dissertation?
Five to fifteen is the normal range, and what you argue for is saturation rather than a number. Write out how you judged that no new themes were appearing.
What is the difference between validity and reliability?
Validity is whether you measure what you intend to measure; reliability is whether the measurement repeats. A scale that consistently reads two kilos light has high reliability and low validity.
Do I need ethical approval for a student dissertation?
Student projects are usually covered by a departmental ethics process rather than a full research ethics committee, but the requirements on consent and personal data still apply. If your study involves sensitive personal data, patients or children, clear it with your supervisor before collecting anything.
Can I change method halfway through?
Yes, if the question turns out to be unanswerable with the original design, but the change should be agreed with your supervisor and reported in the methods chapter. A documented change of course is a methodological choice; a hidden one is a credibility problem.