How to write a methodology section
On this page
The methodology (or methods) section tells readers exactly how you produced your findings. Its two purposes are transparency, so others can judge whether your conclusions are credible, and replicability, so a competent researcher could repeat the study. A strong methods section is precise, ordered and justified; a weak one is vague in the places reviewers care most about.
Methods vs methodology
In many journals the section is simply called "Methods" and describes what was done. In theses and some social science papers, a "Methodology" chapter also explains the philosophical and theoretical reasoning behind the approach: why a qualitative, quantitative or mixed design is appropriate for the question. Check what your department or journal expects. If in doubt, describe what you did and briefly justify each major choice.
The core subsections
Most empirical papers use some version of the following order. Use subheadings; they make the section easy to scan.
1. Research design
State the overall design in the first sentence: a randomized experiment, a cross-sectional survey, a two-wave panel, a comparative case study, a grounded theory interview study, a secondary analysis of administrative data, and so on. Then explain why it fits your research question. For example, a time-lagged design reduces common-method bias compared with a single survey, and interviews suit questions about how people experience a process.
2. Participants, sample or data sources
Describe who or what was studied, and how they were selected:
- the population and the sampling frame
- the sampling method (random, stratified, convenience, purposive, snowball)
- inclusion and exclusion criteria
- the sample size and how it was determined (power analysis, saturation, available data)
- response rates and attrition for surveys or panels
- key demographic or contextual characteristics (often summarized in a table)
For secondary data, name the dataset, its provider, years and unit of analysis, and how you accessed it.
3. Materials, instruments and measures
For each key variable or concept, say how it was measured. In quantitative work, name the scale, give an example item, the response format, how scores were computed and reliability evidence (for example, Cronbach's alpha or omega in your sample) with a citation to the original instrument. In qualitative work, describe the interview guide or observation protocol and how it was developed or piloted.
Do not forget control variables and how they were chosen; reviewers often ask why a variable was included or left out.
4. Procedure
Describe what happened, in chronological order: recruitment, consent, data collection steps, timing between waves, incentives, and any manipulation or intervention. Someone reading this subsection should be able to picture the study from a participant's point of view.
5. Data analysis
Explain how you analysed the data, linked to each research question or hypothesis:
- data cleaning, missing data handling and exclusions
- the main statistical models or qualitative analysis approach (for example, structural equation modelling, multilevel regression, thematic analysis, process tracing)
- how mediation, moderation or robustness were tested
- software and versions
- in qualitative research, coding procedures, how themes were developed, and steps taken for trustworthiness (memoing, double coding, member checking)
A short table mapping each question to its variables and analysis is often clearer than paragraphs.
6. Validity, reliability and limitations of the design
Briefly note the main threats to validity and how the design addresses them: common-method variance, selection bias, confounding, researcher bias. Save the full discussion of limitations for the discussion section, but show here that you anticipated the problems.
7. Ethics
State ethics approval (committee and reference number, if applicable), informed consent, anonymization and data storage. For sensitive topics or vulnerable groups, explain additional safeguards.
How much detail is enough?
The test is replicability: could a competent researcher in your field repeat the study? Include anything that could change the results if done differently. Leave out routine steps that any researcher would do the same way, and move long material (full questionnaires, interview guides, extra tables) to an appendix or supplementary file.
Writing style
- Use past tense for what you did: "Participants completed…", "We estimated…".
- Use active voice where it is clearer; many fields now accept "we".
- Be concrete about numbers: dates, durations, sample sizes, response options.
- Justify non-obvious choices briefly with a citation to methodological literature.
- Keep results out: descriptive statistics of the sample can appear here, but findings belong in the results.
Writing the methods before data collection
If you are writing a proposal or drafting before your data exist, write the methods in the future tense (for a proposal) or as a planned design, and mark anything not yet known as an explicit placeholder: [insert final N], [insert alpha], [insert response rate]. This lets you finish the structure early without inventing numbers you might forget to replace. See how to write a research proposal for the planning version of this section.
Common mistakes
- Naming a design without explaining why it suits the question.
- Missing sample size justification or response rate.
- Measures described without sources, items or reliability.
- An analysis plan that does not match the hypotheses.
- Mixing results or interpretation into the methods.
- Overlong descriptions of software clicks instead of decisions.
Quick checklist
- Is the design stated in the first sentence of the section?
- Are sampling, inclusion criteria and sample size justified?
- Is every variable or concept in your research questions measured, with a source?
- Could someone reconstruct the procedure step by step?
- Is each hypothesis matched to a specific analysis?
- Are ethics and data handling addressed?
Once the methods are settled, the rest of the paper becomes easier to write: the results follow the analysis plan, and the abstract can summarize the design in one sentence. See how to structure a research paper for how the methods fit into the whole.