How to Write a Strong Dissertation Research Question

Most dissertations wobble long before the first chapter is drafted. The reading is done, the plan looks tidy, yet the question underneath it all is too broad, or unanswerable, or not really an argument. A research question is not a title you polish at the end of the project. It is the tool you use to decide what to read, what to leave out, which evidence counts and when to stop. Fix it early and everything after it — the literature review, the method, the chapter structure — becomes noticeably easier to control.

A topic is not a research question

'Nineteenth-century women's writing' is a topic. 'Food poverty in the UK' is a topic. Topics are areas; questions are enquiries with a direction of travel. The difference matters, because a topic can absorb almost anything you throw at it, while a question tells you when the work is finished.

A useful question usually contains three things: a specific subject, an angle of interest running through it, and at least two plausible answers. If only one answer is possible, you have written a description. If dozens are possible, the question is too loose to answer within your word count.

Narrow the subject before you shape the question

Broad topics rarely shrink on their own. You have to squeeze them, and there are three directions in which to do it.

1. Who or what you are studying — a named group, text, organisation, site or dataset.
2. Where and when — a period, region, institution or moment that gives the study boundaries.
3. Which angle — the mechanism, relationship, tension or comparison you want to examine.

Take 'social media and young people'. Adding a group (first-year undergraduates at one university), a setting and time (in England, since 2020) and an angle (how private group chats shape a sense of belonging) turns a shapeless topic into something a single dissertation can address. Add too much, though, and you create a question nobody could answer in a year. Narrowing is a balance, not a competition.

Four tests every dissertation question should pass

• Focused. Could you answer this within your word count, your timescale and your access to evidence? If it would take three separate studies, cut it back.
• Arguable. Is there a case to be made, and a counter-case worth answering? Questions beginning 'What are the main factors…' often invite a list rather than an argument. Ask how, why or to what extent instead.
• Feasible. Can you obtain the sources, participants or materials you need? Are there ethical or access barriers you cannot clear? A superb question you cannot research is not a superb question.
• Significant. So what? Who would care about the answer, and why does it matter to your field? You need not change the world, but you do need a reason beyond personal curiosity.

Weak questions and stronger rewrites

Compare a common starting point with a version that could actually be researched. 'Is remote working good for productivity?' is broad, invites a yes-or-no answer and hides a pile of assumptions about what productivity means. A stronger version might ask: 'How do team leaders in small software firms describe changes in collaboration after moving to hybrid working?' It names a group, a setting and a particular kind of change, and the answer will be an argument about patterns rather than a verdict.

Was the British Empire profitable? That is a library. How did mid-Victorian railway investors write about risk in India? That is a dissertation.

Notice what the stronger versions do. They use verbs that commit to something — describe, shape, constrain, account for, change — and they leave room for the evidence to disagree with you. A question that cannot be answered wrongly is not worth asking.

Get the wording right

Small changes in phrasing shift what counts as a good answer.

• How and why questions suit explanations and processes.
• To what extent suits comparisons and degrees of effect.
• What role does X play in Y suits relational questions in politics, sociology and media studies.
• Closed yes-or-no questions are risky unless you can genuinely argue both sides and your reader expects a verdict.

Keep the question to one sentence. Two-part questions joined by 'and' usually become two dissertations in one, and you will end up answering the easier half properly and the other half badly. Use the precise terminology your field uses, but resist jargon that would collapse under a plain-English explanation.

Expect it to change

Your first version is a working question, not a vow. After twenty sources or one pilot interview, the sharpest question often shifts by a few degrees. That is normal, and usually a sign of good reading rather than failure. The point is to make the change deliberate: note what you learned, explain how the question responds to it, and check that the new version still passes the four tests. Supervisors tend to be far more comfortable with a question that has evolved for stated reasons than with one that has stayed vague and untouched for six months.

A practical sequence to follow

1. Write your topic in one line, then jot down five questions about it without editing them.
2. Circle the one you would genuinely enjoy answering. Sustained interest carries a dissertation further than ambition does.
3. Apply the three filters — who, where and when, which angle — and write the question again.
4. Run the four tests. If one fails, adjust the question rather than the test.
5. Draft a one-paragraph answer to your own question as if you already knew the result. If you cannot, the question may be unanswerable as written.
6. Take both versions to your supervisor and ask which is more researchable, not which sounds more impressive.

Then keep the question visible. Put it at the top of your notes, in your proposal and above the plan for every chapter. A dissertation that answers one clear question, even imperfectly, reads far better than one that circles several. Decide what you are asking, and let everything else serve it.

Photo: paulorear / Pixabay