How to Write a Research Methodology Chapter

Introduction

Of every chapter in a thesis or dissertation, the methodology chapter is the one examiners scrutinise most closely. It is the chapter where a vague phrase like “a survey was conducted” triggers a string of follow-up questions: Who was surveyed? How many? How were they chosen? What instrument was used and was it validated? How were the data analysed? What ethical safeguards were in place? A scholar who cannot answer these questions defensibly has not yet finished the chapter — no matter how many words are on the page.

The good news is that a strong research methodology chapter is not a creative-writing exercise. It is a structured, defensible account of the decisions you made and the reasons you made them. Examiners and reviewers are not looking for originality here; they are looking for transparency, justification, and alignment between your research questions, your design, and your analysis plan. Once you understand the structure that examiners expect, writing the chapter becomes a matter of methodically working through each required component.

This guide lays out a complete framework for writing a defensible methodology chapter — from research design through ethical considerations — with the named frameworks that examiners recognise. If you would like a PhD-qualified specialist to review your draft chapter, our data analysis and methodology service provides section-by-section feedback within 48 hours. You may also find our companion guides on writing research methodology, the data analysis chapter, and choosing between quantitative and qualitative research useful.

What Is a Research Methodology Chapter?

A research methodology chapter is the section of a thesis, dissertation, or research proposal that explains how the study was conducted and why each methodological choice was made. It is distinct from the methods section of a journal article, which is typically a compressed 500–1,000-word summary. In a thesis, the methodology chapter is usually 4,000–8,000 words and is expected to be exhaustive.

The chapter must answer two parallel questions. First, what did you do? — the procedural account of design, sampling, instrument, data collection, and analysis. Second, why did you do it that way? — the methodological justification that links each decision back to your research questions, your epistemological stance, and the literature. The first question is descriptive; the second is argumentative. A chapter that answers only the first reads like a lab manual. A chapter that answers only the second reads like a philosophy essay. Examiners want both.

Saunders, Lewis, and Thornhill’s “research onion” (2019) remains the most widely taught framework for structuring this chapter. It moves outward from philosophy (positivism, interpretivism, pragmatism) through approach (deductive, inductive, abductive), strategy (experiment, survey, case study, ethnography), choice (mono, mixed), time horizon (cross-sectional, longitudinal), and finally techniques (sampling, instrument, analysis). Each layer justifies the next.

The Five Pillars of a Methodology Chapter

A defensible research methodology chapter rests on five interlocking components. Omit any one and the chapter is incomplete.

1. Research design — the overall plan that links your research questions to the data you will collect. Quantitative (experimental, quasi-experimental, correlational, survey), qualitative (phenomenology, grounded theory, ethnography, case study), or mixed-methods.
2. Population and sampling — who or what you studied, how you selected them, and how many. This is the section most frequently queried by examiners; vague statements here sink otherwise strong theses.
3. Data collection — the instrument (survey, interview schedule, observation protocol, archive), its source (adopted, adapted, or self-developed), and its validity and reliability evidence.
4. Data analysis plan — the techniques you used, the software (SPSS, R, NVivo, MAXQDA), and the assumption checks or trustworthiness criteria you applied.
5. Ethical considerations — IRB approval, informed consent, confidentiality, data protection, and any specific risks to vulnerable participants.

Each pillar deserves its own H2 or H3 section. Do not collapse them into a single paragraph; the visible structure is itself evidence of methodological rigour.

Writing the Research Design Section

Open the chapter by restating your research questions or hypotheses and naming the design you chose. If you ran a quasi-experimental pre-test/post-test study, say so explicitly. If you conducted a phenomenological study using Moustakas’s method, name it. Vague phrases like “a qualitative approach was used” are not enough; an examiner needs to know which qualitative approach and why.

Justify the design with reference to your research questions. A “how” or “why” question calls for qualitative or case-study designs. A “how many” or “is there a difference” question calls for quantitative or experimental designs. If your questions span both, mixed-methods is the natural choice — but only if you justify the specific variant (convergent, explanatory sequential, exploratory sequential) with reference to Tashakkori and Teddlie’s typology. For more on the design decision, see our paradigm comparison guide.

State your epistemological position briefly. A positivist stance supports hypothesis testing and statistical generalisation; a constructivist stance supports inductive theory-building and transferable insights. Pragmatism underwrites mixed-methods research. You do not need to write a philosophy essay — one focused paragraph is sufficient — but the stance must be visible because it governs how the rest of the chapter is read.

Population, Sampling, and Instrumentation

This is the section where theses most often fail. Be exhaustive. Define the target population precisely, with explicit inclusion and exclusion criteria. Name the sampling frame (the actual list or platform from which you drew the sample) and acknowledge any mismatch between frame and population. State the sampling technique (simple random, stratified, cluster, purposive, snowball) and justify it with reference to your research questions, not convenience.

Report sample size as a calculated number, not a round figure. For surveys of proportions, use Cochran’s formula (n0 = Z²pq/e²) or the Krejcie and Morgan (1970) table. For experimental designs, use G*Power with alpha = 0.05, power = 0.80, and a defensible effect size estimate. For qualitative studies, justify sample size by saturation, not statistics. Our sampling methods guide walks through each calculation with worked examples.

For the instrument, identify whether it was adopted, adapted, or self-developed. Adopted instruments need a citation and a note on permissions. Adapted instruments need a description of what was changed and a re-validation plan (pilot with 30–50 respondents, Cronbach’s alpha ≥ 0.70). Self-developed instruments need the most extensive validation: expert review, cognitive pre-testing, pilot, and reliability analysis. For survey design specifically, see our survey design guide.

Writing the Data Analysis Plan

The data analysis plan must be specific enough that another researcher could reproduce it from your description alone. State the software (SPSS version 27, R 4.3 with the tidyverse, NVivo 14), the procedures (independent-samples t-test, one-way ANOVA, multiple linear regression, thematic analysis), and the assumption checks (Shapiro-Wilk for normality, Levene’s for homogeneity of variance, multicollinearity diagnostics for regression).

For quantitative analysis, name the test, name the variables, and report the alpha level you used to declare significance (conventionally 0.05). If you ran multiple tests, address the multiple-comparisons problem — either by using a correction (Bonferroni, Holm) or by justifying why no correction was needed. Effect sizes are no longer optional: report Cohen’s d, eta-squared, or R² alongside every p-value. Our statistical tests guide and SPSS output interpretation guide cover the full workflow.

For qualitative analysis, name the analytic method (Braun and Clarke’s six-phase thematic analysis, Charmaz’s constructivist grounded theory, Smith’s interpretative phenomenological analysis). Describe the coding process: how codes were generated (inductive, deductive, or hybrid), how inter-coder reliability was established (if multiple coders were used), and how themes were reviewed and refined. State the software used (NVivo, ATLAS.ti, MAXQDA, Dedoose) and address the four trustworthiness criteria from Lincoln and Guba (1985): credibility, transferability, dependability, and confirmability.

Trustworthiness and Validity

Examiners look for an explicit section on the validity and reliability of your findings. Quantitative researchers address internal validity (does the design support causal claims?), external validity (can findings generalise beyond the sample?), construct validity (does the instrument measure what it claims?), and reliability (would the instrument produce the same result on repetition?). For each, name the specific technique you used: random assignment for internal validity, probability sampling for external validity, factor analysis for construct validity, Cronbach’s alpha for reliability.

Qualitative researchers address the four trustworthiness criteria. Credibility is supported by prolonged engagement, persistent observation, and member checking. Transferability is supported by thick description of context. Dependability is supported by an audit trail of decisions. Confirmability is supported by triangulation and reflexivity. A chapter that names these techniques demonstrates methodological literacy; one that simply asserts “the findings are trustworthy” does not.

Ethical Considerations

Every research involving human participants, organisations, or sensitive data requires ethical clearance from a recognised institutional ethics committee (IRB in the US, IEC in India, university ethics board in the UK and Europe). State the approval number and date. Examiners in regulated fields (medicine, psychology, education with minors) will check this number; absence of it can halt a viva.

Describe informed consent: how participants were briefed, what they consented to, and how consent was recorded (signed form, clickwrap, oral). Address confidentiality and anonymity — the two are different. Confidentiality means you know who participants are but will not disclose; anonymity means you do not know who they are. If you collected personally identifiable data, describe the data-protection measures (encrypted storage, password-protected files, data retention schedule) and reference the applicable regulation (GDPR, India’s DPDP Act 2023, HIPAA).

Address specific risks. Vulnerable populations (minors, prisoners, cognitively impaired individuals) require additional safeguards. Sensitive topics (trauma, mental health, illegal behaviour) may require referral to support services. Pay attention to compensation: payments large enough to constitute undue inducement are themselves an ethical problem.

Examiners do not reject methodology chapters for using imperfect methods. They reject them for hiding the imperfections. An honestly reported convenience sample with a stated threat to external validity is more defensible than a disguised one.

Common Mistakes That Get Methodology Chapters Rejected

Across hundreds of methodology chapters we have reviewed, five mistakes account for the majority of major-revision decisions:

• Method-research question mismatch. The design does not actually answer the stated questions. A phenomenological question cannot be answered with a survey; a prevalence question cannot be answered with 15 interviews.
• Vague sampling reports. Phrases like “a representative sample was drawn” without naming the technique, frame, or size calculation. Examiners want the Cochran or G*Power calculation, not an assertion.
• Unvalidated instruments. A self-developed questionnaire used without pilot testing or reliability analysis. Even adapted instruments need a re-validation note.
• Missing assumption checks. Parametric tests (t-test, ANOVA, regression) reported without normality or homogeneity tests. Reviewers increasingly reject papers that skip these checks.
• Uncited methodological frameworks. Use of Braun and Clarke’s thematic analysis or Creswell’s design typology without citation. Always cite the source of any framework you adopt.

Run your draft against this checklist before submission. Fixing these issues takes hours; receiving a major-revision decision takes months. Our editing and proofreading service includes a methodology-specific review that catches these issues before you submit.

Conclusion

A research methodology chapter is the most defensible chapter in your thesis if you treat it as a structured, transparent account of decisions and justifications rather than a creative-writing exercise. Restate your research questions, name your design and epistemological stance, then walk the examiner through population and sampling, instrument and validation, data analysis plan with assumption checks, trustworthiness or validity evidence, and ethical safeguards. Cite the named frameworks examiners recognise — Saunders’ research onion, Creswell’s design typology, Braun and Clarke’s six phases, Lincoln and Guba’s trustworthiness criteria, Cochran’s sample-size formula. Avoid the five common mistakes and your chapter will pass examiner scrutiny without major revisions. If you would like an expert review, our data analysis team offers chapter-level consultations. Pair this guide with our methodology writing primer, data analysis chapter template, and SPSS walkthrough, and reach us through our contact page for a free consultation.

Frequently Asked Questions