A pillar guide on how to write a methodology section for a research paper that survives peer review — design, sampling, instruments, analysis and the justifications reviewers expect in SCI, SSCI and Scopus journals.
The methodology section is where reviewers spend the most time deciding whether your paper deserves further consideration. The introduction may set expectations, the discussion may interpret findings — but it is in the methods section that reviewers form their hardest judgement: can I trust this study?
This guide walks through how to write a methodology section for a research paper that reviewers will respect: from research design to sampling to measurement to analysis. The principles apply whether you are submitting to an SCI engineering journal, an SSCI management journal, or a Scopus-indexed education journal.
Strip the methodology section down to its essence, and reviewers are asking four questions:
Many otherwise-strong papers fail on the second of these. Writing “we used PLS-SEM” is not a methodology choice; it is an assertion. Writing “we used PLS-SEM because the model is exploratory and the sample size (n = 187) is below the threshold typically recommended for covariance-based SEM (Hair et al., 2022)” — that is a methodology choice. The justification, not the technique, is what convinces reviewers.
The first methodological choice is not which test to run — it is which paradigm fits your question. The wrong fit leads to reviewer comments that no amount of polishing can fix.
| Dimension | Quantitative | Qualitative | Mixed |
|---|---|---|---|
| Best for | Testing hypotheses, measuring effects | Exploring processes, meanings, lived experience | Triangulating evidence, sequential explanation |
| Typical data | Surveys, experiments, archival numbers | Interviews, observation, documents | A combination, in distinct phases |
| Typical analysis | Regression, SEM, ANOVA, ML models | Thematic, content, grounded theory | Phased — quant then qual, or vice versa |
| Sample size | Often 150–500+ | Often 15–40 informants | Depends on phase weighting |
| Common in | SCI, SSCI management, psychology | SSCI education, sociology, AHCI | Education, health, policy research |
Within each, our companion guides go deeper: quantitative methodology — what to report and how, qualitative methodology — what reviewers expect, and mixed methods research — how to justify and structure it.
Most empirical methodology sections in indexed journals follow a similar internal structure. Reviewers expect each of these elements in roughly this order. Missing one will trigger reviewer comments.
State and justify your design (cross-sectional survey, quasi-experiment, multiple case study, mixed methods sequential explanatory, etc.). One paragraph.
Describe the context (organisation type, country, industry, demographic) and inclusion or exclusion criteria. Specific enough that a reader can judge generalisability.
State the sampling technique (random, stratified, purposive, snowball) and justify the sample size — power analysis for quant, saturation logic for qual.
How the data were actually obtained: survey distribution, interview protocol, lab procedure, archival extraction. Include dates, response rates, and any incentives.
For each construct, cite the original scale, the number of items, the response format, and reliability statistics (Cronbach's α, composite reliability).
The statistical or analytical procedures applied — software used, model specifications, any robustness checks. This should track the hypotheses one-by-one.
IRB approval, informed consent, anonymity protections. Increasingly mandatory across SCI/SSCI journals — see our guide on how to write an ethics statement.
Different fields weight these differently. An SCI engineering paper may compress sampling and amplify experimental procedure. An SSCI qualitative paper may expand setting and participants but compress instruments. The seven elements are still expected — only the proportions shift.
Every methodological choice should be justified, and the strongest justifications cite either a methodological authority (Hair, Creswell, Yin) or precedent in the literature you are entering. “Following [authoritative source]…” and “consistent with prior studies in this domain (citations)…” are the two most useful constructions.
Three places where justification matters most:
If you have not yet drafted your methodology, this sequence keeps you from the most common traps. Each step is about decision-making before writing — not about prose.
The method serves the question, not the reverse. If the question is still vague, finalise the introduction before drafting methods.
Quantitative, qualitative or mixed. Half-committing produces half-convincing methods sections.
For every H1, H2, H3, write the specific test you will use. If a hypothesis has no analysis, drop it or split the paper.
Every scale needs a citation trail. Original source, recent validations, reliability statistics from your sample.
“Participants were recruited…”, “Data were analysed using…”. Most reviewers expect this convention.
Appropriate? Justified? Replicable? Defensible? If you cannot answer yes to each, the section is not finished.
The methodology section is technical, but it is not a lab notebook. Two opposite mistakes are common: too vague to be replicable, or so detailed that the argument is buried. The right balance reports every detail a competent peer would need to reproduce the study — and nothing else.
A useful test: when you describe your data collection, could a doctoral student in your sub-field replicate it from your description alone? If yes, your methodology has the right level of specificity. If they would need to email you for clarification, add detail.
A methodology section is judged not on what it claims, but on what it lets a sceptical reviewer verify.
Method without justification. Naming a technique without explaining why. The reviewer will ask why not the alternative? — and not finding an answer triggers revision.
Sample without size justification. A sample of 87 with no power calculation, or a sample of 850 with no rationale for that scale. Both invite scrutiny.
Self-developed instruments without validation. If you built a scale yourself, you need a validation procedure — pilot test, expert review, exploratory factor analysis. Otherwise reviewers will not trust your measurement.
Vague analytical procedures. “We analysed the data using SPSS” is not a method description. Specify the test, the model specification, and the software version.
Missing ethics statement. Increasingly non-negotiable across SCI, SSCI and Scopus journals. IRB approval number, informed consent procedure, and anonymity protections.
Misaligned with the framework. Constructs in the methodology that did not appear in the literature review, or hypothesised paths that have no corresponding test. Reviewers cross-check.
Unlike the discussion section, where prose flow matters, the methodology benefits from clear sub-headings: Research Design, Participants, Sampling, Measures, Analysis. Reviewers can locate exactly what they want to evaluate — and a navigable section is rated more favourably than a perfectly written but unstructured one.
The skeleton is the same; the muscle differs by field. Knowing how to write a methodology section for your specific journal means matching the conventions of your sub-area.
Heavy on experimental setup, dataset specifications, model architecture, benchmark comparisons, and evaluation metrics. Code and data availability statements expected. Reporting requirements have tightened since 2024 — see reproducibility in CS research.
Heavy on construct validity, measurement model, confirmatory factor analysis, and common-method bias checks. SEM-based papers should match the reporting standard documented in SEM and CFA reporting.
Heavy on positionality, interview protocol, coding procedure, trustworthiness criteria (credibility, transferability, dependability, confirmability), and data saturation. Quotes are evidence — but the procedure for selecting them needs to be transparent.
Reporting follows checklists — CONSORT for trials, STROBE for observational studies, PRISMA for systematic reviews. Methodology sections track these checklists almost item by item.
A methodology section that lives in isolation is half a section. It must connect backward to the literature review (constructs match, framework aligns) and forward to the results (every analysis stated is run, every hypothesis tested). For more on those connections, see how to structure a research paper for Scopus/WoS journals.
This integration is what turned a strong concept into a Springer Nature publication for a research team we worked with — the methodology was not stronger than the rest of the paper; it was more tightly aligned with the rest of the paper.
Before submission, run your methodology section through these questions. They mirror what a strong reviewer will check first.
Knowing how to write a methodology section for a research paper is one of the highest-return investments early-career researchers can make. Strong methods rescue weaker writing; weak methods sink even elegant prose. If you would like a senior editor to audit your methodology — for design appropriateness, justification depth, and reporting completeness — that is exactly what our manuscript preparation service covers.
Upload your draft to the AI Manuscript Health Checker. Get scores on design justification, reporting completeness, and analytical alignment — plus reviewer-style suggestions for each.
Check Methodology Score Book a Free Consultation