How do I structure a PhD thesis chapter by chapter?

The generic shape exists because it answers a reader's questions in the order they arise. What are you asking and why does it matter. What is already known. How did you find out. What did you find. What does it mean. What follows. Every discipline's variation is a rearrangement of those six moves.

Where theses differ most is the middle. A three-study thesis in psychology has three parallel empirical chapters, each with its own short method and results. A humanities thesis may have four analytical chapters with no separate methodology section at all. A thesis by publication has chapters that are papers, plus connective material. All are correct in their own fields, and none transfers.

So the reliable move is to read three recent, well-regarded theses from your own department, ideally ones supervised by your supervisor. Look at the table of contents rather than the prose. That tells you what your examiners expect, which is the only structural question that matters at submission.

Then agree the structure with your supervisor rather than deciding it alone. The advice in the PhD-completion literature is that the supervisor advises on the thesis structure and on the number and length of chapters, that student and supervisor settle an overall plan about one year before submission, and that they review that plan every few weeks (Marino et al., 2014, p. 1). Note the timing. The plan is a document you revise repeatedly in the final year, not a decision you make once.

One decision worth making early: what each empirical chapter is defined by. One sub-question per chapter is the version readers follow most easily, because each chapter has a question at the top and an answer at the bottom. Chapters defined by data source or by method tend to sprawl, since nothing tells you when they are finished.

Order those chapters by how the evidence supports itself. Zhang's rule for research papers is to arrange experiments and findings according to their argumentative relationships, so that earlier results establish the premises later results depend on, and to cut arguments that repeat and dilute the reader's attention (Zhang, 2014, p. 1). The same source suggests a fast planning method: sketch the figures and tables first, put them in the order the argument needs, make each one readable on its own, and only then draft the prose (Zhang, 2014, p. 1). A chapter sequence that survives that test rarely needs restructuring later.

How long should each chapter be?

In an 80,000 word thesis, roughly: introduction 6,000, literature review 15,000, methodology 10,000, three empirical chapters at 12,000 each, discussion 8,000, conclusion 5,000. Treat those as proportions rather than targets, and check your own regulations for the word limit that binds.

The two chapters that most often run over are the literature review and the methodology, for the same reason: both are easier to write than the analysis, so they absorb the anxiety of not yet having results. A literature review at 30,000 words is usually a signal that the empirical work is behind.

Check whether your word limit includes footnotes, references, and appendices, since the answer varies and the difference can be several thousand words.

If you are writing an MA or MSc thesis, scale the whole structure down rather than copying a doctoral one. An MSc project usually runs six months to one year and mostly applies and extends coursework, while a PhD runs three to four years in most European countries, with a reported median of six to ten years in the USA (Saccenti and Furlan, 2025, pp. 1-2). That difference argues for a tighter question, fewer empirical chapters, and a shorter discussion at master's level. Two related warnings sit alongside it. Saccenti and Furlan name "procrastination by computation," running more analysis to postpone writing (Saccenti and Furlan, 2025, p. 5), and Marino advises starting to write once the agreed experiments are done instead of returning to the laboratory for more data (Marino et al., 2014, p. 2).

When should I write the introduction?

Draft it early so you have a statement of what the thesis claims, then rewrite it last, when you know what the thesis actually claims. The final introduction is written from the finished chapters and almost never resembles the first version.

The early draft is not wasted. It functions as a plan, it gives your supervisor something to react to, and rereading it after two years tells you exactly how far the project has moved.

Whichever draft you are on, give the introduction real work to do. It should state the problem, review the relevant literature, define the research question or hypothesis, and explain how the later chapters answer it, with the literature review sitting inside the introduction and often reworked afterwards into a standalone review article (Marino et al., 2014, p. 3). That places the review differently from the separate literature review chapter many departments expect, so check your own regulations before you follow it.

Write the conclusion before the final introduction. Once you know what you concluded, the introduction's job becomes obvious: set up precisely the question the conclusion answers, and nothing else.

Keep a one-page outline of the whole thesis somewhere you see it often, with each chapter's job in a sentence. It is the fastest way to notice that a chapter has drifted away from what it was supposed to do.