Write a methodology that a reviewer can follow

Explain the logic, not just the method name

A methodology section should show how the proposed evidence will answer the research question. Naming a survey, experiment, interview study or machine-learning model is not enough. Explain why the design fits the question and what its limits are. A reviewer should be able to follow the path from the problem to the planned conclusions.

Begin with the unit of analysis and the main comparison or interpretation. Clarify whether you are estimating an effect, exploring experiences, describing a pattern or evaluating a system. Different purposes require different evidence and different claims.

Describe the data and access route

State what data or participants you need, where they would come from and how access would be obtained. Distinguish a public dataset from a dataset that requires permission. For participants, explain the intended setting and eligibility rather than claiming a sample is already recruited.

Include a reasoned sampling or sample-size plan appropriate to the design. Do not choose a number only because it looks substantial. Where a precise calculation is not yet possible, state the assumptions and the work needed to refine it. Avoid claiming representative coverage when the proposed recruitment cannot support that claim.

Make the procedure reproducible in principle

Outline the major stages in the order they would happen: preparation, permissions, recruitment or data acquisition, measurement, quality checks and analysis. Identify important instruments or materials and whether you need permission to use them. Explain how data would be recorded, stored and checked.

For a computational project, describe the evaluation design, baselines, separation of training and testing where relevant, and how you would prevent avoidable leakage. For qualitative work, explain how records would be produced and how interpretation would be documented. Use discipline-appropriate detail instead of a generic paragraph copied between designs.

Address ethics and limitations early

Identify privacy, consent, risk, fairness, access and confidentiality issues that apply to the project. State which institutional review or permission would be sought. Do not write as though ethics approval has already been granted unless it has.

Explain the main threats to the conclusions and how you would reduce them. Limitations are not an admission that the project is worthless; they define what the design can reasonably establish. Keep the proposed claims proportionate to the evidence.

Turn the method into a work plan

Map tasks to realistic phases and include dependencies. Analysis cannot begin before data are available; recruitment may depend on review approval; an instrument may need piloting. Add time for revision and unexpected delays.

Finish with a consistency check: does every method serve the question, does every claimed outcome have a corresponding analysis, and can the work be completed with the resources described? Remove impressive-sounding methods that do not have a clear role. A coherent methodology is more persuasive than a long inventory of techniques.

Practical editorial guidance, not a guarantee of an outcome. The current programme and provider instructions take precedence.