The methods section of a research paper is the easiest section to write badly and the hardest to fake. Reviewers may skim your introduction and forgive a clumsy transition in the discussion, but they read the methods with a checklist in their heads, because this is where a study earns or loses its credibility. If a competent colleague could not repeat your work from what you wrote, the section has failed—no matter how polished the prose.
The good news is that methods sections are the most formulaic part of academic writing. Conventions are stable, readers expect a predictable order, and originality is actively discouraged. This guide walks through what the section must accomplish, the components to include, how expectations shift across research types, the style conventions editors enforce, and the mistakes that draw reviewer fire.
What the Methods Section of a Research Paper Must Accomplish
Within the familiar IMRaD structure, each section answers one question. The introduction answers why, the results answer what, the discussion answers so what—and the methods answer how. Concretely, the section has three jobs:
- Enable replication. A trained reader in your field should be able to repeat the study, or at least evaluate exactly what was done, from your description alone.
- Justify your choices. Readers need to see that the design fits the question. Brief rationale for key decisions belongs here, not a full defense.
- Expose limitations honestly. Sample size, instruments, and procedures carry constraints. Stating them plainly in the methods builds the trust your discussion section will spend.
Keep the audience in view: you are writing for a skeptical peer, not a student. Skip textbook explanations of standard techniques, cite them instead, and spend your words on what was specific to this study.
The Core Components, in the Order Readers Expect
Design
Open with a sentence naming the overall design: a randomized experiment, a cross-sectional survey, a comparative case study, an ethnography, a corpus analysis. This orients everything that follows. If your design responds directly to your research question, one clause of rationale here pays off—readers who want help sharpening that link can start with our guide to writing a research question.
Participants, Sample, or Sources
Describe who or what you studied and how you selected them: recruitment, inclusion and exclusion criteria, sample size, and relevant characteristics. In humanities and archival work, this becomes a description of your corpus, archive, or primary sources and the logic of their selection. The selection logic matters as much as the count.
Materials and Instruments
Name your instruments, measures, apparatus, or interview protocols with enough precision that a reader could obtain or reconstruct them. For established instruments, cite the source and report any adaptation you made. For instruments you built, summarize their development and include the full version in an appendix or supplementary file.
Procedure
Narrate what happened, in order, from the participant’s or researcher’s point of view: what was done, by whom, under what conditions, over what period. This is where vague drafts collapse. “Data were collected over several months” tells a reviewer nothing; “each participant completed two 45-minute interviews, four weeks apart, in their workplace” can be evaluated and repeated.
Analysis
Close by explaining how raw data became findings: statistical tests and software, coding procedures and codebook development, analytic framework, or interpretive approach. Name the decisions a skeptic would ask about—how you handled missing data, how disagreements between coders were resolved, why this test rather than the obvious alternative.
How Expectations Differ by Research Type
The underlying contract—say what you did so others can judge it—holds everywhere, but the details readers demand vary by tradition. The table below is a quick orientation for the three broad families most academic authors work in.
| Element | Quantitative | Qualitative | Humanities / Archival |
|---|---|---|---|
| Sample or sources | Size, power, sampling method | Purposeful selection, saturation logic | Corpus or archive and selection rationale |
| Instruments | Validated measures, reliability | Interview or observation protocols | Editions, translations, provenance |
| Analysis | Named tests, software, assumptions | Coding process, framework, reflexivity | Theoretical lens, interpretive method |
| Replicability ideal | Exact repetition | Transparent, auditable process | Traceable evidence and reasoning |
Mixed-methods papers describe each strand separately, then explain how the strands were integrated—concurrently or sequentially, and with which strand given priority. The integration sentence is the one mixed-methods reviewers look for first.
Tense, Voice, and Style Conventions
Editors enforce a few near-universal conventions. Write the section in past tense: you are reporting what was done, not issuing instructions. Most style guides now welcome first-person active voice (“we randomized participants”) over strained passive constructions, though field norms vary—check two or three recent articles in your target venue and match them.
Use subheadings once the section passes a few hundred words; reviewers navigate methods by subheading. Report numbers consistently, define abbreviations at first use, and resist the urge to editorialize. Sentences like “this innovative design” belong nowhere in a methods section—the design should look rigorous on its own evidence.
Finally, state ethics plainly where human participants are involved: review board approval, informed consent, and how confidentiality was protected. One or two matter-of-fact sentences are standard, and their absence is one of the fastest ways to earn a revise-and-resubmit.
Common Mistakes in the Methods Section of a Research Paper
- Writing a recipe with no rationale. Steps without reasons make sound decisions look arbitrary. One justifying clause at each major choice is enough.
- Smuggling in results. Response rates and sample descriptions belong in methods; findings do not. If a sentence answers the research question, move it.
- Under-describing the analysis. “Themes emerged from the data” and “standard statistical procedures were used” are the two phrases reviewers mock most. Name the process.
- Over-describing the standard. Two paragraphs explaining what a t-test is signals inexperience. Cite and move on.
- Quiet inconsistencies. A sample of 120 in the methods and 117 in the results table, with no explanation, invites suspicion of everything else. Account for every discrepancy.
Transparency Extras: Data, Materials, and Preregistration
Expectations around research transparency have risen across disciplines, and the methods section is where you signal that you have met them. None of these practices is universally required, but each one strengthens a submission where it applies:
- Data availability. Say where your data, corpus, or transcripts can be accessed, or state honestly why they cannot be shared, such as confidentiality agreements with participants.
- Materials and code. Posting instruments, stimuli, or analysis scripts in a repository lets reviewers verify details you had no room to print, and a one-line pointer in the methods costs nothing.
- Preregistration. In fields where hypotheses are tested, noting that the design and analysis plan were registered before data collection, with a link, defuses concerns about selective reporting.
- Deviations. If the study drifted from the original plan, say what changed and why. Reviewers forgive documented changes far more readily than discovered ones.
Treat these as trust accelerants. A reviewer who can open your protocol or rerun your script spends less energy doubting you and more energy engaging with the work itself.
A Drafting Process That Prevents the Usual Problems
Draft the methods section while the study is running, not months later from memory; contemporaneous notes capture the small procedural details reviewers ask about, from software versions to the exact wording of a recruitment email. Then revise against a simple test: hand the draft to a colleague in your field and ask them to describe, step by step, what you did. Every point where they hesitate or guess is a gap in the writing, not a failure of their reading, and each one is cheap to fix before submission and expensive after.
Remember, too, that the methods section does its work again in the discussion, where every limitation you acknowledge should trace back to something disclosed here—our guide to the discussion section of a research paper shows how the two sections mirror each other. Writing them as a pair, rather than months apart, keeps the paper honest and internally consistent.
Conference versions of a paper are an excellent stress test. Presenting your design to a mixed audience surfaces the assumptions you forgot to state, and peer-reviewed venues that publish full papers—including NIVA’s official conference proceedings and NIVA journals—give you a refereed round of methods scrutiny before the journal submission that matters most.
Frequently Asked Questions
Methods answer what you did; the introduction has to explain why anyone should care. See how to write the introduction of a research paper.
Typically 500 to 1,000 words, or roughly 10 to 20 percent of the paper, though field norms vary widely. The real standard is sufficiency: long enough for a trained peer to evaluate and repeat the work, and no longer.
Past tense, because you are reporting completed actions: participants were recruited, interviews were conducted, data were analyzed. Present tense appears only for statements that remain true, such as describing where an archive is held.
Yes, though it may be titled differently, such as approach, sources, or methodology. Readers still need to know what you examined, how you selected it, and how you analyzed it, so they can judge whether your conclusions follow from your process.
Methods are the specific procedures you used: the survey, the coding scheme, the statistical test. Methodology is the broader framework justifying those choices. Journal sections usually want methods, with just enough methodological rationale to show the design fits the question.
A methods section sharpens fastest under real questions from real peers. Bring your study to NIVA’s next virtual conference and put your design in front of reviewers and colleagues across twenty-plus disciplines—faculty registration is $150, students pay $70, and joining NIVA is free.