Start with a question, not a topic

Almost every dissertation that goes wrong in the final weeks went wrong in the first fortnight, and the usual culprit is a topic that was never narrow enough to be answerable. "Something about social media and mental health" is a topic. "How do first-year undergraduates in the UK describe the effect of late-night social media use on their sleep?" is a question, and a question is something you can actually research.

Aim for a working question you can state in one sentence, ideally under thirty words. If you cannot, you are probably still holding two or three questions at once. Write them all down, then choose the one that excites you enough to spend nine months with it. Boredom with your own title is a genuine risk; so is picking something so fashionable that three other students in your seminar are doing the same thing.

Good dissertation questions tend to have three qualities. They are specific enough to be answered within your word count, contestable enough that a reasonable person could disagree, and resourced enough that the evidence exists and you can reach it.

Test the question before you commit

Before you write a single chapter heading, run your question through a short feasibility check. It takes an afternoon and saves months.

• Scope: can you answer it in 8,000–12,000 words? If your answer would need a book, cut a variable. If it could be answered in a paragraph, you have a question for an essay, not a dissertation.
• Sources: do a quick library search. Are there at least twenty credible academic sources you can name now? If not, refine the question rather than hoping material appears later.
• Access: if you need interviews, participants, archives or equipment, can you realistically get them? Find out about ethics approval timelines in week one, not week ten.
• Method: can you say, roughly, how you will gather and analyse evidence? A question with no plausible method is a wish, not a plan.

Write your answers in a single page of notes and show it to your supervisor. That page will generate a far more useful meeting than arriving with a vague idea and hoping for direction.

Build the chapter skeleton before you read widely

Many students read for months and then try to squeeze their notes into chapters. Reverse that order. A provisional chapter plan gives your reading a purpose, because every source you open can be filed against a chapter that needs it.

Most dissertations in UK universities settle into a recognisable shape, though your discipline may vary:

• Introduction: the question, why it matters, how the dissertation is structured.
• Literature review: what is already known, where the gaps and disagreements lie.
• Methodology: your approach, your evidence, your ethical considerations and limitations.
• Findings or analysis: often two or three chapters, each handling one theme or dataset.
• Discussion: what your findings mean in relation to the literature.
• Conclusion: the answer, its limits, and where research could go next.

Write each chapter heading as a short sentence rather than a single noun. "Chapter 3: How participants describe the pressure to be constantly available" tells you what to write. "Chapter 3: Findings" tells you nothing and invites you to dump everything you have.

Give every chapter a job and a word count

An outline becomes genuinely useful once each chapter has a purpose, an approximate length, and a sentence describing what it must achieve. Allocate word counts early, even roughly. If your dissertation is 10,000 words, a plan might look like 800 for the introduction, 2,500 for the literature review, 1,200 for methodology, 3,500 across two analysis chapters, and 2,000 shared between discussion and conclusion.

Two things follow from this. First, you can see immediately whether a chapter is overburdened; if your literature review needs 5,000 words to make sense, your question is probably too broad. Second, you gain a set of small, finishable tasks. "Write 900 words on how the concept is defined in the literature" is far easier to start on a Tuesday morning than "write my dissertation".

Under each heading, add three to five bullet points of the argument you expect to make. These are not commitments; they are hypotheses about your own writing that will change as you read.

Let the plan move as your reading does

A chapter outline is a working document, not a contract. Expect to revise it after your first serious block of reading, again after your methodology is settled, and once more when your findings turn out to be different from what you assumed. Save each version with a date so you can see how your thinking developed — supervisors often find that evolution interesting, and it can be worth a paragraph in your introduction.

Revise for a reason, though. Adding a chapter because you found an interesting article is usually a mistake; changing a chapter because your evidence does not support your original claim is exactly right. Keep a running note of decisions you have made and why, so you are not relitigating the same choices in month six.

Take the first practical steps this week

Planning is only useful if it produces action. Within the next seven days, try to do four things:

• Draft your question in one sentence and a paragraph explaining why it matters.
• Sketch the six or seven chapter headings with approximate word counts.
• List ten sources you already know are relevant, and note which chapter each belongs to.
• Send the one-page plan to your supervisor and ask specifically whether the scope is realistic.

None of this requires you to have read everything or to know your answer. It requires only that you decide what you are trying to find out and where in your dissertation you will say it. Do that, and the blank page becomes a set of manageable sections — which is all any dissertation really is.