How to Write a Complete Methods Section
A practical guide to writing the statistical methods section of a manuscript — covering every element peer reviewers and reporting guidelines require.
You'll learn
How to write a complete, reproducible Methods section that satisfies peer reviewers and reporting guidelines.
Use this when
You are drafting the Methods section of any manuscript.
Why the Methods Section Is Critical
The Methods section enables reproducibility. A well-written Methods section allows another researcher to re-run your analysis and arrive at the same numbers. If a peer reviewer cannot understand exactly what you did, your paper will be rejected or require major revisions.
💡 Write Methods before you write Results
Writing the Methods section before you draft your Results forces you to commit to your analysis plan and reduces post-hoc analytical decisions. It also ensures you have not omitted any steps when the analysis is fresh in your mind.
Required Elements — Statistical Methods Paragraph
- 1.Study design in one sentence: "This was a retrospective cohort study..."
- 2.Population and eligibility: who was included and excluded, and how many
- 3.Primary and secondary outcomes: defined precisely (how measured, at what timepoint)
- 4.Exposure/predictor variables: how they were defined and categorized
- 5.Statistical tests: name every test used — not "appropriate tests" but "two-sample t-test," "Chi-square test," "Cox proportional hazards regression"
- 6.Software and version: "Analyses were performed using R 4.3 (R Foundation for Statistical Computing)"
- 7.Significance threshold: "Statistical significance was defined as p < 0.05 (two-tailed)"
- 8.Missing data handling: state the approach and whether assumptions were assessed
- 9.Sensitivity analyses: pre-specified sensitivity analyses that were planned
Sentence Templates That Satisfy Reviewers
- ●Descriptive: "Continuous variables are expressed as mean ± SD or median [IQR] based on normality assessed by the Shapiro-Wilk test."
- ●Comparison: "Between-group differences were compared using the independent samples t-test for normally distributed variables and the Mann-Whitney U test for non-normal variables. Categorical variables were compared using the Chi-square test or Fisher exact test as appropriate."
- ●Regression: "Multivariable logistic regression was performed with variables selected a priori based on clinical relevance. Results are expressed as adjusted odds ratios (aOR) with 95% confidence intervals (CI)."
- ●Missing data: "Missing data were handled using complete case analysis. Sensitivity analyses using multiple imputation by chained equations (MICE) were performed to assess the impact of missing data assumptions."
Trusted sources behind this lesson
Further reading
Read next
Machine Learning for Researchers: A Practical Introduction
Understand the ML workflow — training, validation, testing — and how it differs from traditional statistical hypothesis testing. No coding required.