Research Methods for a Dissertation: A Guide

What a methodology chapter is — and what it must justify

The methodology chapter (often Chapter 3 of a UK dissertation) is the part of your thesis where you explain how you conducted your research and, crucially, why you did it that way. It is not a passive description of what you did; it is an argument. A strong methodology persuades the examiner that your chosen methods were the most appropriate available for answering your research questions, and that the data they produced can be trusted.

To do that, your methodology has to justify a connected set of decisions. Every one of them should trace back to your aims and research questions:

• Research philosophy — your assumptions about what counts as valid knowledge (positivism, interpretivism, pragmatism, critical realism).
• Research approach — whether you reason from theory to data (deductive) or from data to theory (inductive), or combine both (abductive).
• Methodological choice — qualitative, quantitative, or mixed methods.
• Research design / strategy — survey, experiment, case study, ethnography, grounded theory, action research, and so on.
• Data-collection methods — questionnaires, interviews, observation, secondary data, documents.
• Sampling strategy — who or what you studied, how you selected them, and how many.
• Data-analysis methods — statistical tests, thematic analysis, content analysis, and so on.
• Quality and ethics — reliability, validity (or trustworthiness), and ethical safeguards.

A useful way to picture how these decisions nest inside one another is Saunders, Lewis and Thornhill’s “research onion“, where you peel from the outer layer (philosophy) inwards through approach, strategy, choices and time horizons to the core (data collection and analysis). You do not have to use the onion explicitly, but the logic — outer assumptions constrain inner methods — is sound and worth internalising.

“The research onion provides an effective progression through which the methodology for a piece of research can be designed… its underlying logic is that the outer layers must be addressed before the more central, data-collection issues.” (Source: Saunders, Lewis & Thornhill, Research Methods for Business Students)

The big choices, layer by layer

1. Research philosophy and approach (inductive vs deductive)

Your philosophy sets the rules for what you treat as evidence. Positivism assumes an objective, measurable reality and favours quantitative testing of hypotheses. Interpretivism assumes reality is socially constructed and favours rich, qualitative understanding of meaning. Pragmatism sits between them and says: use whatever methods best answer the question — the natural home of mixed methods.

Your approach follows from this. A deductive approach starts with existing theory, derives a hypothesis, and collects data to test it — “theory first”. An inductive approach starts with observations and builds theory or themes up from the data — “data first”. Many real dissertations are abductive, moving back and forth. For a fuller treatment of this distinction, see our guide to inductive and deductive reasoning.

2. Qualitative vs quantitative vs mixed methods

This is the choice students agonise over most, so it deserves care. Quantitative research measures variables numerically and tests relationships statistically; it answers “how much”, “how many”, and “does X affect Y”. Qualitative research explores meaning, experience and process through words and observation; it answers “how” and “why”. Mixed methods deliberately combine the two — for example, a survey to map a pattern followed by interviews to explain it.

The decision is not about which is “better” but which fits your question. A question about the prevalence of student burnout is quantitative; a question about how students experience burnout is qualitative. Our dedicated comparison of quantitative vs qualitative research goes deeper, and the table below summarises the trade-offs.

3. Research design (strategy)

Your design is the overall plan that links philosophy and approach to concrete data. Common designs include:

• Survey — collecting standardised data from many respondents to describe or compare.
• Experiment — manipulating an independent variable under controlled conditions to establish cause and effect.
• Correlational design — measuring whether and how strongly variables move together, without manipulation.
• Case study — an in-depth investigation of one or a few bounded cases in real-world context (Yin’s work is the standard reference here).
• Ethnography / grounded theory / action research — immersive or iterative qualitative strategies.

The right design depends on the kind of claim you want to make. If you need to claim causation, you need an experimental or quasi-experimental design; observation alone can only establish association. The types of research guide maps these strategies in more detail.

4. Data-collection methods

Once the design is set, you choose the instruments that actually gather data. Primary methods generate new data (questionnaires, interviews, focus groups, observation, experiments); secondary methods reuse existing data (official statistics, archives, prior datasets, documents). Most dissertations use one or two methods well rather than many superficially. Our overview of the methods of data collection compares each instrument’s strengths, costs and pitfalls.

5. Sampling

You almost never study an entire population, so you sample from it. The headline split is between probability sampling (random, systematic, stratified, cluster) — which supports statistical generalisation — and non-probability sampling (purposive, convenience, snowball, quota) — which is common, defensible and often necessary in qualitative work. What matters is that your sampling logic matches your aim: representativeness for quantitative generalisation, information-richness for qualitative depth. See the full sampling methods of research guide before you commit.

6. Data-analysis methods

Your analysis must match your data type. Quantitative data is analysed with descriptive statistics (means, frequencies, dispersion) and inferential statistics (t-tests, ANOVA, correlation, regression, chi-square) to test hypotheses. Qualitative data is analysed with techniques such as thematic analysis (Braun & Clarke’s 2006 six-phase approach is the most cited framework), content analysis, or discourse analysis. The cardinal rule: state the specific technique and justify it — “the data were analysed” is not a method.

7. Reliability, validity and ethics

Reliability is consistency — would the same method produce the same results again? Validity is accuracy — are you measuring what you claim to measure, and are your conclusions warranted? In qualitative work these are reframed as trustworthiness (credibility, transferability, dependability, confirmability, per Lincoln & Guba). You should also address ethics: informed consent, confidentiality, the right to withdraw, data protection, and ethics-committee approval. Our guide to reliability and validity explains the threats and how to mitigate them.

How to choose your methods: align everything to your research questions

The single most important principle in methodology is alignment (sometimes called methodological congruence): your philosophy, approach, design, data-collection, sampling and analysis should all point in the same direction, and that direction is set by your research questions. Examiners look for this thread; broken alignment — for instance, an interpretivist framing followed by a hypothesis test — is one of the most common reasons methodology chapters lose marks.

Work through these questions in order:

1. What exactly am I trying to find out? Write the research question as precisely as possible. Its grammar reveals the method: “how many / to what extent / does X predict Y” signals quantitative; “how / why / what is the experience of” signals qualitative.
2. What kind of knowledge claim do I need to make? Causation needs an experiment; prevalence needs a representative survey; deep understanding needs qualitative depth.
3. What does my discipline expect? Lab psychology leans experimental; education and sociology often lean qualitative or mixed; business and health frequently use surveys and mixed methods.
4. What is feasible? Honestly assess time, access, sample availability, skills and ethics. A perfect design you cannot execute is worse than a good design you can.
5. Does every layer agree? Check that philosophy → approach → design → collection → sampling → analysis form one consistent chain.

Recommended structure of the methodology chapter

While supervisors vary, a robust dissertation methodology chapter typically follows this numbered structure. Use it as a checklist and adapt the headings to your field:

1. Introduction — restate the aim and research questions and signpost the chapter.
2. Research philosophy — state and justify your paradigm (positivist, interpretivist, pragmatist).
3. Research approach — inductive, deductive or abductive, and why.
4. Research design / strategy — survey, experiment, case study, etc., with justification.
5. Data-collection methods — the instruments, how you developed them, and any piloting.
6. Sampling — population, sampling technique, sample size and recruitment.
7. Data-analysis methods — the specific techniques and software (SPSS, R, NVivo) and how you applied them.
8. Reliability and validity (or trustworthiness) — how you safeguarded quality.
9. Ethical considerations — consent, confidentiality, approval, data protection.
10. Limitations of the methodology — honest acknowledgement of constraints.
11. Chapter summary — a short recap linking the methods back to the questions.

For where this chapter sits within the whole thesis and how it connects to your literature review and findings, see our complete guide on how to write a dissertation.

A worked example: writing a methodology paragraph

Knowing the components is one thing; writing them in flowing, justified prose is another. The example below shows what a single justified paragraph from a methods chapter actually reads like — notice how every sentence does work and links a choice to a reason.

That single worked paragraph names the philosophy and approach, the design, the sampling, the instruments, the analysis technique and software, and the quality and ethics safeguards — every methodological layer, each tied to a reason. It is a model you can adapt: swap in your own paradigm, design, sample, methods and analysis, and keep the same justify-as-you-go density throughout the chapter.

Strengths and limitations: be honest

No method is flawless, and examiners reward critical awareness. Quantitative designs offer breadth, replicability and statistical power but can miss context and meaning. Qualitative designs offer depth and nuance but limit generalisation and are more exposed to researcher bias. Mixed methods can offset each other’s weaknesses but demand more time, skills and careful integration. State your design’s limitations explicitly and explain how you mitigated them — doing so strengthens, rather than weakens, your chapter.

Common mistakes to avoid

• Describing without justifying. Saying what you did but never why is the most common failing. Every choice needs a reason tied to the question.
• Misalignment. An interpretivist framing with hypothesis testing, or a tiny convenience sample used to claim generalisation.
• Vague analysis. “The data were analysed thematically” without naming the framework or steps.
• Confusing methodology with methods. Methodology is the rationale and logic; methods are the specific techniques. The chapter needs both.
• Ignoring ethics and limitations. Omitting consent, data protection or honest constraints reads as naivety.
• Over-claiming causation. Correlational or survey data cannot prove cause; choose your verbs carefully.

How to do it well: a quick checklist

1. State your research questions, then justify every method against them.
2. Make the philosophy → approach → design → collection → sampling → analysis chain explicit and consistent.
3. Name specific techniques, instruments, software and frameworks — no vagueness.
4. Address reliability/validity (or trustworthiness) and ethics in their own sections.
5. Acknowledge limitations honestly and explain your mitigations.
6. Cite methodological authorities (Saunders, Creswell, Bryman, Yin, Braun & Clarke) to anchor your choices.

Get these right and your methodology becomes the most defensible chapter in your dissertation rather than the most feared one.

Related methodology guides

• The Research Onion (Saunders)
• Research Philosophy
• Mixed Methods Research

Frequently Asked Questions