How do I write a methodology chapter?

The methodology chapter is where doctoral candidates most often lose marks for something they actually did well. The research design was sound, the choices were deliberate, and the chapter reports them as a sequence of events rather than a set of decisions, so the reader cannot see the judgment.

Write it as a chain of decisions. Each one has a form: here is the choice I faced, here is what I chose, here is the alternative, here is why. Four sentences per decision, ten to fifteen decisions in a typical chapter, and the reader can now see a researcher thinking rather than a procedure being followed.

Start drafting the chapter while the design is still movable. Zhang argues that an outline of the eventual paper belongs in the design stage, because it forces you to state your objectives, fix the logic of the investigation, and organize your data before you collect it, and because writing then works as a test of the project rather than a report on it. That test can expose a weak design, a broken chain of reasoning, or results that do not support the claim, and the fix is revised methods or more data (Zhang, 2014, p. 1). A methodology chapter written after the analysis is finished can only describe what happened.

Include the decisions that went wrong. A pilot that failed, a recruitment strategy that did not work, a measure you abandoned. These read as weaknesses to a nervous candidate and as evidence of real research to an examiner, and concealing them means the examiner finds the seam and wonders what else is missing.

Say something about your philosophical position if your field expects it, and keep it short and functional. A page connecting your epistemological stance to your actual design is useful. Five pages of general philosophy that never touch your study is a chapter an examiner will skip and then ask about.

How much methodological detail is enough?

Enough that a competent researcher in your field could repeat the study and get comparable results. That means instruments, sampling, procedure, and analysis in the chapter, with instruments and full protocols in appendices. If a detail would change the result, it belongs in the text.

Reporting standards for your design are the fastest way to check coverage. Most fields have a checklist for common designs, and running your chapter against one takes half an hour and catches omissions reliably.

Kass and colleagues put a usable floor under "enough detail": someone holding your data and your description of the analysis should be able to recreate your tables, figures, and statistical inferences. They also note the barriers that break this in practice, including different software versions, different settings, and different computing environments, which is why naming a package is not a description of an analysis (Kass et al., 2016, p. 6). The same paper warns that software and algorithms are tools rather than explanations, so state why a method answers your substantive question instead of citing the procedure and moving on (Kass et al., 2016, p. 4).

Qualitative work meets that bar differently, and the template shifts with it. Korstjens and Moser report that qualitative articles typically run 5,000 to 7,000 words and need thick description of participants and context, transparent reflection on how the methods were applied, and fuller treatment of findings than a quantitative paper usually carries. They also note that the standard Introduction, Methods, Results, Discussion order is not compulsory, and that ethnographic work may open with a narrative or a case, or merge findings and discussion (Korstjens and Moser, 2018, p. 123). Keep a structure the reader can follow, and change its proportions to fit the design.

The details most often missing are the boring ones: exactly how participants were recruited, how many declined, what the response rate was, and how missing data was handled. Those are the ones a careful reader wants.

How do I justify my choices instead of just describing them?

For each decision, name the alternative you did not take and say why. "Semi-structured rather than structured interviews, because the concepts were not yet stable enough to fix the questions" is a justification. "Semi-structured interviews were used" is a description an examiner will question.

The reason should come from your research question rather than from convenience or convention. "Because that is standard in this field" is weak on its own, and becomes strong when you add why the standard fits your question.

Your justification for quality checks has to use the criteria that fit your design. Korstjens and Moser name credibility, transferability, dependability, and confirmability for qualitative research, with reflexivity part of what makes the work transparent, and they state that internal validity, generalizability, reliability, and objectivity are not suitable standards for judging it (Korstjens and Moser, 2018, pp. 120-121). Their triangulation example is specific enough to copy the form of: in-depth interviews, focus groups, and field notes, several researchers involved, two of them analysing the first six interviews independently before comparing interpretations, meetings after every third data set, and regular team analysis sessions (Korstjens and Moser, 2018, p. 122). Describe checks at that level of detail only where they actually happened in your study.

Where the honest reason was practical, say so plainly. "A larger sample was not feasible within the funded period, which limits the precision of the estimate" is a defensible sentence. An invented methodological justification for a practical constraint is not, and examiners recognize it.