Key takeaways
- Write a dissertation in stages: topic, questions, proposal, literature review, methods, data, results, discussion, then introduction, conclusion and abstract.
- Start writing in week one. Notes, memos and rough methods text all become draft material later.
- Write the introduction and abstract last, once you know what you actually found.
- Plan backwards from your deadline and protect at least three to four weeks for revision and editing.
- Agree a feedback rhythm with your supervisor or advisor early, and send them short, specific drafts rather than whole chapters at the last minute.
How do you write a dissertation?
You write a dissertation by moving through a sequence of manageable stages: choose a focused topic, turn it into research questions, plan your method in a proposal, review the literature, collect and analyse your data, then write up results, discussion, introduction, conclusion and abstract, in roughly that order. Most of the difficulty comes not from any single stage but from leaving writing until the research is “finished”.
The steps below apply equally to a master's thesis, an undergraduate dissertation or a doctoral project. The scale changes; the logic does not.
Step 1–2: choose a topic and define your questions
Choose a topic you can actually finish. A good topic sits where three things overlap: something you are curious about, a genuine gap or debate in the literature, and data or sources you can realistically access in the time you have. Our guide to choosing a dissertation topic includes a shortlisting exercise.
Turn the topic into research questions. A topic is an area (“student wellbeing and part-time work”); a research question is something your study can answer (“Is the number of weekly paid working hours associated with self-reported burnout among full-time undergraduates in their final year?”). Good questions are specific, answerable with your method and neither trivially obvious nor impossibly broad. One main question with two or three sub-questions is a common, workable shape. See our guide on how to write research questions for examples.
Write your questions on a single page and keep them visible. Every later decision, from which articles to read to which tables to include, should trace back to them.
Step 3: write the proposal
Most programmes require a research proposal or prospectus before you begin data collection. Even when it is not assessed, writing one is worth it, because it forces the key decisions early. A typical proposal covers:
- the problem and why it matters;
- your research questions or hypotheses;
- a short literature review that shows the gap;
- your design, sample or sources, and analysis plan;
- ethical considerations and approvals needed;
- a timeline.
Do not treat the proposal as a hurdle to clear and forget. A good proposal becomes the skeleton of your first three chapters. Reused and expanded, it can give you a running start of several thousand words.
If your project involves human participants, apply for ethics approval (an IRB in the US) as soon as your proposal is accepted. Approval can take weeks, and you cannot collect data without it.
Step 4–5: review the literature and plan the method
Literature review. Search systematically: decide on databases and search terms, record what you searched and when, and keep every source in a reference manager (Zotero, Mendeley, EndNote) from day one. Read for argument, not just content. For each key source, note its question, method, main finding and limitation in two or three lines. When you write, organise the review by theme or debate, not by author. A paragraph that begins “Smith (2019) found…” followed by “Jones (2021) found…” is a summary; a paragraph that begins with a claim and uses several studies as evidence is a review.
Methodology. Your method must match your question. Questions about how much, how often or whether X predicts Y usually call for quantitative designs. Questions about how people experience, understand or negotiate something usually call for qualitative designs. Mixed methods work when each strand answers a different part of the question. Write the methodology chapter in draft before you collect data. It doubles as your operating manual, and it is far easier to write while decisions are fresh.
A good methodology chapter justifies its choices as well as describing them. For each major decision (design, sampling, instrument, analysis), add a sentence explaining why it suits your question and what the realistic alternative was. For example: “A cross-sectional survey was chosen over a longitudinal design because the study aimed to estimate associations within a single academic year; the limitation this imposes on causal inference is discussed in Chapter 5.” Examiners look for exactly this kind of reasoning, and it prepares you for the questions you will be asked in your viva or defense.
Want a second pair of eyes? Want a clear plan and expert feedback on each chapter while you write your dissertation yourself? Request a free quote for dissertation help. See dissertation help.
Step 6–7: collect, analyse and report your results
Keep a research log during data collection: what you did each day, problems you encountered and decisions you made. This log will write half of your limitations section for you.
For analysis, follow the plan you set out in your proposal and record any deviation with a reason. If you are working with statistics, check your assumptions, report effect sizes and confidence intervals alongside p-values, and keep a script or syntax file so every result can be reproduced. If you are working qualitatively, document how codes were developed and how themes were refined.
Then write the results (or findings) chapter. Report what you found, organised around your research questions, without interpreting it yet. Tables and figures should be readable on their own, with clear titles and labelled units. A short, plain sentence often does more than a paragraph:
Students working more than 15 hours per week reported higher burnout scores (M = 3.41, SD = 0.72) than those working 15 hours or fewer (M = 2.96, SD = 0.69), t(212) = 4.62, p < .001, d = 0.64.
If your analysis is the part that worries you most, our dissertation statistics help walks you through each test so that you can run it, report it and defend it yourself.
Step 8–9: discussion, introduction, conclusion and abstract
Discussion. This is where you earn most of your marks. For each research question, state the answer, compare it with the studies from your literature review, offer plausible explanations, and be honest about limitations. Avoid repeating numbers from the results chapter. Interpret them instead. A simple paragraph pattern works well: finding (what you found, in words), comparison (which earlier studies it agrees or disagrees with), explanation (why that might be) and implication (what follows for theory or practice).
Introduction. Now rewrite the introduction from your proposal. You know what the dissertation actually argues, so the introduction can promise exactly what it delivers. End it with a short roadmap of the chapters.
Conclusion. Answer the main question directly, summarise your contribution, give practical or theoretical implications and suggest specific further research. Keep it shorter than the discussion and do not introduce new findings.
Abstract. Write it last, in one sitting, as a compressed version of the whole: context, aim, method, key results and main conclusion. Our abstract guide includes a worked example.
A realistic dissertation timeline
The timeline below assumes a nine-month master's project. Compress or stretch the proportions for your own deadline.
| Months | Main focus | Writing output |
|---|---|---|
| 1 | Topic, questions, initial reading | One-page project summary; reading notes |
| 2 | Proposal, ethics application | Proposal (2,000–4,000 words) |
| 3–4 | Literature review; finalise instruments | Draft literature review; draft methodology |
| 5–6 | Data collection and analysis | Research log; analysis outputs; results tables |
| 7 | Write results and discussion | Full draft of results and discussion |
| 8 | Introduction, conclusion, abstract; supervisor feedback | Complete first draft |
| 9 | Revision, editing, formatting, references | Final submission |
Notice that writing appears in every month. If data collection slips, as it often does, you can still progress the review and methods.
Practical habits that make it easier
- Write little and often. Three hundred words a day for five days a week is 6,000 words a month. A short daily target beats an occasional marathon.
- Separate drafting from editing. Get the argument down first; polish sentences later.
- Use headings as a plan. Draft each chapter's subheadings before writing, then fill them in. It turns a blank page into a list of small tasks.
- Ask for targeted feedback. Instead of “any comments?”, ask your supervisor “Does section 2.3 make the gap clear enough?”
- Reference as you go. Inserting citations at the end is slow and error-prone.
- Back up everything in at least two places, including your data and analysis files.
- Leave time to edit. A final read for consistency, formatting and referencing (APA, Harvard or your required style) catches problems that would otherwise cost marks.
If you would like regular accountability and expert feedback without handing over the writing, dissertation coaching gives you a structured plan and a reader who knows what examiners expect.