Writing the Methodology Chapter: Design, Sampling and Justification

The research onion in practice, philosophy, approach, strategy, sampling, analysis and ethics, with the justification pattern markers reward.

The methodology chapter is graded on one skill above all: justification. Describing what you did earns a pass. Explaining why it was the right choice for your research questions, and why you rejected the alternatives, earns distinction. Structure the chapter as a series of justified decisions, from abstract to concrete.

1. Research Philosophy

State your epistemological position and connect it to the nature of your question. Positivism treats reality as objective and measurable, natural fit for hypothesis-testing, quantitative work. Interpretivism holds that social reality is constructed through meaning, fit for exploring experiences and perceptions. Pattern to follow: "This study adopts an interpretivist stance because its object, how nurses experience workload pressure, exists in participants' meanings rather than in externally measurable facts (Saunders, Lewis and Thornhill, 2019)."

2. Approach: Deduction vs Induction

Deductive research starts from theory and tests hypotheses. Inductive research starts from data and builds toward theory. Your literature review already decided this: if it ended in testable hypotheses, you are deductive. If it ended in open questions about an underexplored phenomenon, inductive.

3. Strategy and Design

Survey: breadth, standardisation, statistical power, at the cost of depth.
Case study: depth and context, at the cost of generalisability. Cite Yin (2018) for design logic.
Experiment: causal inference through control and manipulation.
Ethnography / interviews: lived experience and meaning-making.
Mixed methods: justify the combination (triangulation? explanation of quantitative results?), not just "both is better".

4. Sampling

Name the technique and defend the size. Probability sampling (simple random, stratified, cluster) supports statistical generalisation. Non-probability sampling (purposive, snowball, convenience) is legitimate for qualitative work when justified by information-richness, not ease. For surveys, justify n against a power calculation or accepted rules of thumb. For interviews, invoke saturation, the point where new participants stop yielding new themes (commonly 12–20 for homogeneous groups).

5. Data Collection and Analysis

Describe instruments precisely: questionnaire scales and their sources (adapted from validated instruments where possible, with Cronbach's alpha reported), interview guide structure, pilot testing. Then the analysis pathway: for quantitative, descriptive statistics, tests chosen and their assumptions (e.g. "an independent-samples t-test, following confirmation of normality via Shapiro–Wilk"). For qualitative, name the method and follow it visibly (Braun and Clarke's six-phase thematic analysis is the standard citation).

6. Ethics, Validity, Limitations

Ethics: informed consent, anonymisation, data storage, institutional approval reference.
Quantitative quality: validity and reliability. Qualitative quality: credibility, transferability, dependability, confirmability (Lincoln and Guba, 1985).
Limitations: state them precisely and explain mitigation, honesty here builds marker trust for your findings chapter.

The Justification Formula

[Choice] was selected because [link to research question], consistent with [citation]. Alternatives such as [rejected option] were considered but rejected because [specific weakness for this study].

Apply that sentence pattern to every layer above and the chapter writes itself.

What Goes in the Chapter, in What Order

Departments vary, but the underlying logic runs from the abstract to the concrete, and each section should narrow what came before it.

Section	Question it answers	Rough share
Philosophy or paradigm	What assumptions about knowledge underpin this?	5 to 10 per cent
Approach	Qualitative, quantitative or mixed, and why	10 per cent
Design	Survey, experiment, case study, ethnography?	15 per cent
Population and sampling	Who, how many, selected how?	15 to 20 per cent
Data collection	What instruments, administered how?	20 per cent
Analysis	What you did with the data, and using what	15 to 20 per cent
Rigour	Validity and reliability, or trustworthiness	10 per cent
Ethics	Approval, consent, confidentiality, data handling	5 to 10 per cent

Some departments place philosophy in a separate chapter and some omit it entirely. Check the handbook before writing three thousand words on epistemology that nobody asked for.

Tense, Voice and Person

Choice	Convention	Note
Tense	Past throughout	You are reporting what was done, not proposing it
Voice	Discipline-dependent	Passive still dominates the sciences, active is increasingly accepted
Person	First person acceptable in qualitative work	In reflexive research it is often required

The chapter converts from proposal to dissertation by a tense change, which is why proposal text pasted straight in reads oddly. Search for "will be" before submission.

Replicability Is the Test

The standard the chapter has to meet is straightforward: could a competent stranger repeat your study from this text alone? That question resolves most uncertainty about how much detail to include.

Instruments: name them, cite the source, state how many items and what scale, and reproduce them in an appendix.
Procedure: what participants actually experienced, in order, including timing and setting.
Recruitment: where, when, through what channel, and against what inclusion criteria.
Analysis: the specific tests or analytic steps, and the software with its version.
Decisions: how you handled missing data, outliers and non-response.

That last point is the one most often omitted. Missing data handling is a methodological choice with consequences, and stating it pre-empts an obvious viva question.

Justification Beats Description

A chapter that describes what you did earns a pass. A chapter that justifies it earns a distinction. The difference is visible sentence by sentence.

Descriptive (weak)	Justified (strong)
Semi-structured interviews were used.	Semi-structured interviews were used because the study sought participants' own accounts while retaining comparability across cases, which a fully unstructured format would not have supported.
A questionnaire was distributed online.	Online distribution was chosen to reach students across three campuses within the fieldwork window, accepting the coverage limitation that it excludes those without reliable access.
Thematic analysis was conducted.	Reflexive thematic analysis was selected over content analysis because the question concerned meaning rather than frequency, and over IPA because the sample was heterogeneous.

Notice that each strong version names an alternative and says why it was rejected. That single move demonstrates that a choice was made rather than defaulted into.

The Ethics Section

Short, but its absence is conspicuous. Cover the approval reference, how informed consent was obtained, how confidentiality and anonymity were maintained, how data were stored and for how long, the right to withdraw and its cut-off, and any specific risk with its mitigation. Where your study involved a vulnerable group or a sensitive topic, say more rather than less.

Common Mistakes and Fixes

Mistake	Fix
A textbook essay on research philosophy	State your position and its consequence for your design in a short paragraph
Method described but never justified	Apply the justification formula to every layer
Future tense left from the proposal	Convert to past throughout
Sample size asserted, not justified	Give a power calculation, or the qualitative logic you used
Results creeping into the chapter	Methods says what you did. Findings go in the next chapter
No account of missing data or non-response	State what happened and how you handled it
Instruments not in an appendix	Include them. Replicability depends on it

Discipline Notes

Health and medicine: reporting guidelines from the EQUATOR Network often dictate the structure, and ethics detail is scrutinised closely.
Psychology: a Participants, Materials, Procedure, Analysis structure is conventional and usually expected.
Business and management: philosophy sections are common, and organisational access constraints should be stated openly.
Education: practitioner positionality matters, particularly where you are researching your own setting.
Engineering and computing: the emphasis falls on apparatus, parameters, datasets and evaluation metrics rather than paradigm.

Where Ethical Support Fits

Methodology is the chapter where supervision pays the highest return, because a design flaw found now costs a conversation and found later costs a rerun. Statistics advisers, librarians and ethics officers are all part of the same normal support structure.

What stays yours is the design decisions and the account of what you actually did. If the fieldwork departed from the plan, and it usually does somewhere, report the departure. An honest deviation with its rationale reads as competence, whereas a plan presented as though it ran perfectly invites the question of what else was smoothed over.

Frequently Asked Questions

How long should the methodology chapter be?

Commonly 15 to 20 per cent of the total word count, so roughly 2,000 to 3,000 words in a 15,000-word dissertation. Check your handbook, which may specify.

Do I need a research philosophy section?

Depends entirely on the department. Business and social science programmes often require one. Sciences frequently do not. Ask before writing it.

Can I write in the first person?

In qualitative and reflexive work, usually yes, and sometimes it is expected. In quantitative work, conventions are more conservative. Follow your discipline.

What if my method changed partway through?

Report what you actually did, and explain the change and its reason. Methodological adaptation is normal research practice, and concealing it is the real problem.

Where do pilot studies go?

In the methodology, usually before data collection, describing what the pilot changed. A pilot that changed nothing is worth a sentence saying so.

Should limitations appear in this chapter?

Brief methodological caveats can sit here, but the full limitations discussion normally belongs in the discussion or conclusion. Avoid writing it twice.

Your Next Step Today

Take the justification formula and apply it to the five biggest choices in your chapter: approach, design, sampling, collection and analysis. Five sentences. Any choice where you cannot name a rejected alternative is one you have not yet justified, and that is exactly where the marker will press.

Trusted Sources

University of Southern California, The Methodology, in Organizing Your Social Sciences Research Paper. Accessed 11 August 2026.
University of Southern California, Quantitative Methods. Accessed 11 August 2026.
University of Southern California, Qualitative Methods. Accessed 11 August 2026.
University of Southern California, Limitations of the Study. Accessed 11 August 2026.
Andrade C, Sample Size and its Importance in Research. Open access. Accessed 11 August 2026.
EQUATOR Network, reporting guidelines that shape methods sections in health research. Accessed 11 August 2026.

Methodology chapter structures differ by department. Where this guide and your handbook differ, follow your handbook.