A thesis is a long argument with an empirical spine. Once your data exist, most of it is already determined - what you found constrains what you can claim, which constrains how you frame the problem. That is why the introduction is nearly the last thing to write, not the first.
This guide covers the order to write in, what each chapter has to do, a six-month schedule with checkpoints, how to divide your words, and the five things that most often cost a term.
A full draft to work from
ThesisAI drafts complete academic documents with chapter structure and inline citations verified against the papers they came from, so you start from something to revise rather than a blank page.
Inizia a scrivereWrite Out of Order
The reading order is introduction, literature review, methodology, results, discussion, conclusion. The writing order is not.
- Methodology. Write it while you are doing it. Procedural detail decays fast, and a methods chapter written from memory six months later is always thinner than one written in the week you ran the study.
- Results. Write as soon as the analysis is stable. Report, do not interpret.
- Literature review. You have been reading since the proposal, but write it once you know what your results are about, so it sets up the argument you are actually making.
- Discussion. The intellectual centre. This is where your findings meet that literature.
- Conclusion. Short, and written after the discussion so it does not repeat it.
- Introduction. Last. Now you can promise exactly what the thesis delivers.
- Abstract. Genuinely last, in a single sitting, once nothing else will change.
What Each Chapter Has to Do
| Chapter | Share of words | The question it answers |
|---|---|---|
| Introduction | 8 to 10% | What is the problem, why does it matter, and what will this thesis do about it? |
| Literature review | 25 to 30% | What is known, what is contested, and what is missing that I will supply? |
| Methodology | 15 to 20% | What did I do, why was it the right thing to do, and could someone repeat it? |
| Results | 15 to 20% | What did I find, stated without interpretation? |
| Discussion | 20 to 25% | What does it mean, how does it sit with the literature, and what are the limits? |
| Conclusion | 5 to 8% | What is the answer, what follows from it, and what should happen next? |
For a 15,000 word master's thesis that is roughly 1,400 / 4,200 / 2,600 / 2,600 / 3,300 / 900. Drafts that run badly over almost always do so in the literature review, and the fix is deciding what the review is for rather than cutting evenly.
Each chapter has its own guide: introduction, literature review, methodology, results, discussion and conclusion.
A Six-Month Timeline
This assumes a taught master's thesis of 12,000 to 20,000 words with data collection, and roughly twenty hours a week. Scale the months, not the sequence.
| Month | Main work | Milestone to hit |
|---|---|---|
| 1 | Narrow the topic, read widely, define the gap and the research questions, draft the proposal | Questions approved by your supervisor and ethics application submitted |
| 2 | Systematic reading and note-taking, build the theoretical framework, design the study, pilot the instrument | Ethics approval granted; instrument piloted on three people |
| 3 | Data collection. Write the methodology chapter in parallel, as it happens | Collection complete or nearly; methodology chapter drafted |
| 4 | Analysis. Draft the results chapter. Start the literature review now that you know what it must set up | Results chapter drafted; analysis stable enough not to change |
| 5 | Literature review finished, discussion drafted, conclusion drafted | Full draft of all six chapters, however rough |
| 6 | Introduction, abstract, revision, references, formatting, proofreading | Submitted with a week in hand |
Two notes on the schedule. Ethics approval is the single most common cause of a lost month, so submit it in month one even if the design is not final - amendments are routine, late applications are not. And month six is not padding: formatting, reference cleanup and a proper proofreading pass genuinely take three weeks on a document this size.
Getting Started: Topic, Question, Proposal
Month one decides how hard the other five are. Three things have to happen, in this order.
- A topic you can finish. Feasibility beats interest. The test is access: can you actually reach the people or the data? Our guide to choosing a thesis topic covers narrowing one down.
- A question you can answer. One main question, two or three sub-questions, each answerable with data you can realistically get. See how to write a research question and the related aims and objectives.
- A proposal that commits you. The point is not the document, it is forcing the design decisions early enough to change them cheaply - our thesis proposal guide covers the format.
Working With a Supervisor
Supervision quality varies enormously and you can only control your half of it. What works:
- Send work, not questions. "Here are 2,000 words on X, I am unsure about the framing in section 2" gets a useful response. "How should I approach the literature review?" gets a vague one.
- Agree a meeting rhythm in month one and put it in the calendar for all six months. Monthly is the minimum.
- Send material 48 hours ahead. Nobody reads well in the five minutes before a meeting.
- Write the notes yourself after each meeting and email them back. This is what you will rely on when advice from month two contradicts advice from month four.
- Raise problems early. A supervisor told in week two that recruitment has stalled can help. Told in week ten, they can only sympathise.
The Five Failure Points
- Reading without writing. Months of reading with nothing drafted feels productive and is not. Write a paragraph of synthesis for every few papers, from the first week.
- Waiting for ethics or access. Both take longer than anyone plans for. Start both immediately and design around the possibility of a no.
- Scope creep after data collection. Adding a second research question in month four is the classic way to arrive at month six with two half-finished studies.
- Discussion that repeats results. If your discussion restates numbers without connecting them to the literature, the chapter carrying the most marks is doing nothing. Our discussion guide covers the difference.
- Leaving references to the end. Reconstructing 120 citations in the final week is miserable and produces errors. Use a reference manager from day one and cite as you write.
Writing Habits That Finish Theses
Nothing here is clever, and all of it works better than inspiration.
- Fixed hours, not word targets. Two protected hours a day beats a 1,000-word quota that turns a bad day into a failure.
- Draft ugly. Editing a bad paragraph is a task. Producing a perfect one from nothing is not. Separate the two passes.
- Stop mid-thought. Ending a session in the middle of a sentence makes the next one start instantly.
- Keep a decisions log. One file recording every choice and why. In month five you will need to justify a decision from month two, and the log is where it lives - this is also the raw material for your limitations section.
- Version your files by date and back up to two places. Thesis loss is rare and catastrophic.
- Read your own work aloud once per chapter. It catches more than any spellchecker.
Before You Submit
Leave a week. Check that the questions promised in the introduction are the ones answered in the conclusion, that every table and figure is referred to in the text, that every in-text citation appears in the reference list and vice versa, that the formatting matches the handbook, and that every required declaration - plagiarism, ethics, AI use - is signed and attached. Then prepare for the defence, which is a separate task and starts the day you submit.
FAQs About Writing a Thesis
How long does it take to write a thesis?
Six months of part-time work is realistic for a master's thesis with primary data collection, and three to four months if you are using existing data. A doctoral thesis is three to four years, with the writing concentrated in the final year.
What order should I write the chapters in?
Methodology, results, literature review, discussion, conclusion, introduction, abstract. The introduction goes near the end because only then can it accurately promise what the thesis delivers.
How many words should each chapter be?
Roughly: introduction 8 to 10%, literature review 25 to 30%, methodology 15 to 20%, results 15 to 20%, discussion 20 to 25%, conclusion 5 to 8%. Check your handbook, since some count references and appendices in the total and some do not.
How many sources does a thesis need?
For a master's thesis, typically 40 to 80 cited sources, with more read than cited. The number matters less than coverage: an examiner will notice a missing seminal paper long before they count the list.
What is the difference between a thesis and a dissertation?
It depends on the country. In the UK and much of Europe a dissertation is usually the master's document and a thesis the doctoral one; in the US it is the reverse. Our guide to thesis vs dissertation covers the regional differences.
Can I use AI to write my thesis?
Policies differ and are tightening. Most institutions now permit AI for brainstorming, structuring and language editing while prohibiting submitted text you did not write, and most require disclosure of what you used. Read your own institution's rule first - our guide to using AI ethically in academic writing covers what disclosure looks like.
What do I do if my results are not what I expected?
Report them. A null or contradictory finding is a finding, and explaining why the expected pattern did not appear is exactly the kind of work the discussion chapter exists for. Changing the question to match the data after the fact is the thing that actually damages a thesis.
Write the methodology while you are doing it, get ethics moving in week one, and accept that the introduction belongs at the end. Those three decisions do more for a thesis than any amount of extra reading. Start with choosing a topic you can actually finish.