The methodology chapter has a single job that most drafts miss: justification. Examiners don't read chapter 3 to learn what you did — they read it to judge whether what you did can support the claims you make in chapter 5. Every section below must therefore answer two questions: what did you do, and why was that the right choice against the alternatives?
The nine-section structure
- 1Introduction — restate the research questions this chapter must serve (three lines; anchors everything that follows).
- 2Research philosophy & approach — positivist/interpretivist/pragmatist; deductive or inductive. One honest paragraph each, tied to your questions — not a textbook tour.
- 3Research design — survey, experiment, case study, mixed methods… named, defined, and matched to the questions (types of research design).
- 4Population & [sampling](/learn/sampling-methods-in-research) — population defined precisely, frame acknowledged honestly, method justified, size calculated, response rate anticipated.
- 5Instrumentation — every construct's source scale (author, year, items, anchors), adaptations made and why, plus the full instrument in an appendix.
- 6[Validity & reliability](/blog/reliability-and-validity-explained) — content validity procedure, pilot study results, alpha per construct, and (for SEM studies) the validity thresholds you will apply.
- 7Data collection procedure — mode, period, consent process, response handling; enough detail to replicate.
- 8Data analysis plan — each research question mapped to its statistical test or qualitative procedure, with software named.
- 9Ethical considerations & limitations — ethics approval, consent, anonymisation; methodological limitations stated plainly (they reappear in chapter 5).
Worked example: described vs justified
Described (weak): "A questionnaire survey was used. The sample size was 384 employees selected by convenience sampling. Data was analysed in SPSS." Every sentence is true and none of it defends a single decision.
Justified (strong): "A cross-sectional survey design was adopted because the research questions test relationships among perceptions across a broad population — depth-first designs (interviews, case study) could not support the intended generalisation (RQ1–RQ3). The population comprised employees of IT SMEs in Pune (est. 41,000 per NASSCOM regional data); with no accessible sampling frame, proportional quota sampling by firm size was the strongest feasible method, and its generalisation limits are acknowledged in §3.9. The minimum sample of 381 (95% confidence, ±5%) was inflated to 460 for an anticipated 80% usable-response rate…" Same study — but every choice now carries its why, its alternative, and its consequence.
The alignment table: your cheapest insurance
One table mapping research question → objective → hypothesis → instrument section → analysis technique. It forces you to catch misalignment (a question no instrument item serves; a hypothesis no test addresses) before your examiner does, and it quietly answers half the viva. Put it at the chapter's end or in an appendix.
The proposal's methodology is written in future tense ('data will be collected'); the thesis chapter reports completed work in past tense. Convert systematically — leftover future tense is the most common tell of a hastily-assembled chapter. Passive voice is conventional in most Indian faculties; whichever you choose, stay consistent.
The checklist your chapter must survive
- Could a competent stranger replicate the study from this chapter alone?
- Does every choice name at least one rejected alternative and the reason?
- Is every construct's measurement traceable to a cited source scale?
- Do the analysis plan's tests actually answer the research questions as framed?
- Are the limitations here consistent with what chapter 5 later admits?
Where this chapter sits in the whole document is covered in how to structure a PhD thesis. If your design itself is still unsettled — not just its write-up — that's a design problem to fix first, and it's exactly what research methodology mentoring works through before a word is drafted.
Frequently asked
How long should the methodology chapter be?+
Typically 12–20% of the thesis — roughly 8,000–15,000 words for most Indian PhD theses. Length follows justification depth: a mixed-methods or multi-study design legitimately needs more; padding a simple survey design to look rigorous impresses no one.
Should the methodology chapter be in past or future tense?+
Past tense in the thesis (the work is done); future tense in the proposal/synopsis. When converting proposal text into the chapter, change tense systematically and update any procedure that changed in execution — examiners notice mismatches between the stated and actual method.
What is the difference between research design and research methodology?+
Methodology is the whole reasoned framework — philosophy, approach, methods and their justification. Design is one component: the study's structural plan (survey, experiment, case study). Chapter 3 covers the methodology; the design is usually its third section.
phdguide's mentors are senior academics, former supervisors, statisticians and publication specialists with 25+ years' average experience guiding MBA, MPhil and PhD scholars from topic to viva.
Ethical, compliant guidance: We provide academic support, mentoring, analysis, editing and structuring — not authorship. Your work stays compliant with university policies.