Home/Insights/Methodology reference/Methodology in research
Methodology in research
Method is what you did. Methodology is why that was the right thing to do. Papers are rejected for confusing the two.
The distinction
A method is a procedure: a questionnaire, an assay, a regression. A methodology is the reasoning that connects the research question to the choice of method, and that justifies the inferences drawn from the result.
A methods section that lists procedures without that reasoning answers "what was done" and leaves "why this answers the question" unaddressed. Reviewers read the gap as a design problem even when the procedures are sound.
What a methodology has to establish
That the design can answer the question asked. A cross-sectional survey cannot establish that exposure preceded outcome, however large it is.
That the sample supports the inference. Who was eligible, who was approached, who took part, and who the findings are meant to apply to.
That the measurements mean what they are taken to mean — validity and reliability of instruments, not merely their names.
That the analysis matches the design. Clustered data analysed as independent is a methodology failure, not an analysis detail.
Writing it so it can be repeated
The test is whether a competent reader in the field could repeat the study from the text. That standard decides the level of detail: instrument versions, software versions, settings, and the handling of missing data.
Reporting guidelines exist for exactly this and are design-specific: CONSORT for randomised trials, STROBE for observational studies, PRISMA for systematic reviews, COREQ or SRQR for qualitative work. Writing to the relevant checklist before submission is the cheapest available improvement to a methods section.
In practice
Fix, flag or query.
The distinction that keeps an edit defensible: what an editor may change silently, what must be shown to the author, and what only the author can answer.
- Cannot be fixed later: a design that does not match the question.
- Most common omission: how missing data were handled.
- Check against: the EQUATOR Network checklist for your design.
Reporting standards
- EQUATOR Network, reporting guideline library
- ICMJE Recommendations, on the methods section
Reporting guidelines named here are hosted by the EQUATOR Network. Writing to the checklist for your design before submission is the cheapest available improvement to a methods section.
Related