Choosing a dissertation topic feels daunting, but a good topic is rarely found fully formed — it is built, by narrowing a broad interest into one precise, answerable question that fits your word count, timeframe and the data you can realistically obtain. The students who struggle most are not those who pick an uninteresting subject; they are those who pick a subject that is too wide, has no accessible evidence, or needs approval they were never going to get. This guide walks you through the narrowing process step by step, with the feasibility checks supervisors actually apply and a worked example showing how a vague idea becomes a workable project.
Start where your energy already is
Begin with the modules you genuinely enjoyed and performed well in. You will live with this topic for months, through reading, data collection and several thousand words of writing, so curiosity is a practical asset, not a luxury. Look back at your assignment feedback: topics where markers praised your critical engagement often point towards fertile ground. Also look at what frustrated you. A lecture claim you doubted, a theory that did not match your placement experience, or a gap a lecturer mentioned in passing can all become starting points.
At this stage, stay broad on purpose. Write down three or four interest areas in plain language — “staff wellbeing in nursing”, “social media marketing for small businesses”. Then read around each one: skim recent journal articles and note any live debate where two sides disagree, since a debate gives your literature review structure and your discussion something to contribute to. If an area produces no recent literature at all, treat that as a warning sign; silence usually means the evidence base is too thin to build on.
Turning an area into a researchable question
An interest area becomes a dissertation when it can be phrased as a question with a population, a setting and something measurable or explorable at its heart. Compare “social media and mental health” — an area so vast that thousands of books exist on it — with “How do undergraduate students at one UK university describe the effect of late-night social media use on their sleep and next-day concentration?” The second version tells you who you would study, where, and what you would actually ask them. It can be answered with a small survey or a set of interviews inside one term.
A practical formula helps: [action or relationship] + [population] + [setting] + [time or condition]. “Does flexible rostering (relationship) affect self-reported stress (outcome) among care assistants (population) in residential homes (setting) during the first year after a new rota system (condition)?” Not every question needs numbers — qualitative questions using “how” and “what” are equally valid — but every good question could, in principle, be answered with evidence you describe in a methodology chapter.
Test your draft question: could two reasonable researchers disagree about the answer, can you name the data that would answer it, and could you state the answer in one paragraph once the data were in? If not, it is not yet researchable or still too wide.
The feasibility checks supervisors actually apply
Before falling in love with a question, run the four checks your supervisor and ethics panel will run anyway. First, literature: can you find at least 20 to 30 relevant, credible sources, including recent journal articles? Search your university library properly, not just the open web. Second, data: can you obtain the evidence yourself? Interviews require willing participants; questionnaires require distribution channels; secondary data requires a dataset that actually exists and is accessible to students. Third, ethics: is approval realistic? Projects involving children, patients, vulnerable adults or sensitive topics such as illegal behaviour are not impossible, but they need longer approval routes, gatekeeper permission and sometimes checks you may not complete in time. Ask early, before you commit. Fourth, scope: does the question fit the word count? A 10,000-word dissertation can handle one relationship, one setting and one population comfortably. It cannot compare three countries, four generations and two decades of policy.
Time is the check students forget. Work backwards from your submission date: ethics approval, recruitment and transcription all consume weeks that writing cannot recover. Transcribing one hour of interviews takes most beginners four to six hours. Ten interviews is therefore a week of transcription alone. If your plan needs forty interviews, the plan — not your work ethic — is the problem. Our dissertation proposal help page covers how these feasibility decisions are presented convincingly in the proposal itself, which is where supervisors judge them formally.
Narrowing levers and a worked example
When a topic is too broad, narrow it deliberately using one or two levers rather than hacking at it randomly. The reliable levers are population (new staff rather than all staff), place (one trust, one city, one company), period (since a specific policy or event), sector (independent retailers rather than retail as a whole), and a single variable pair (training and confidence, rather than training, confidence, pay, management and culture all at once). Each lever you apply makes the project more doable and, paradoxically, usually makes the findings more interesting, because depth replaces generality.
Here is the process in action. Starting area: “mental health in the workplace”. Lever one, population: narrow to newly qualified teachers. Lever two, place: narrow to secondary schools in one local authority. Lever three, variable pair: focus on mentoring support and intention to leave the profession. Lever four, period: the first two years after qualifying. The researchable question becomes: “To what extent does formal mentoring support influence newly qualified secondary teachers’ intention to leave the profession within their first two years?” That question supports a focused literature review, a feasible survey or eight to twelve interviews, and a discussion with clear practical implications. Notice that nothing about the original interest was lost — it was concentrated.
For subject-specific inspiration while you narrow, browsing curated lists can help you see the level of precision expected. Nursing students, for instance, will find suitably scoped examples on our nursing dissertation topics page, and business students can compare theirs against the business dissertation topics collection — use them to calibrate scope, then make the question your own.
Supervisor fit, the proposal conversation and topics that get rejected
A supervisor does not need to be the world expert in your exact question, but the fit matters. Supervisors supervise best inside or near their own field: they know the key literature, spot methodological problems early and can suggest sources you would never find alone. Before the proposal conversation, read what your prospective supervisor has published and check the department’s list of supervision areas. In the meeting itself, arrive with two or three draft questions rather than one fixed title, explain your feasibility thinking briefly, and listen carefully to which option draws the most engaged response. Supervisors often steer students towards the version they can support best, and that steer is valuable information, not interference.
Finally, learn from the topics that get rejected, because the reasons repeat every year. Topics fail because no data source exists (“attitudes of chief executives” with no access route), because ethics approval would take longer than the project allows, because the question is a current-affairs essay with no research component (“Is remote working good or bad?”), because the scope needs a research team rather than one student, or because the literature is almost entirely journalistic. A rejection at proposal stage is the system working properly — it is far cheaper to rebuild a topic in week two than to discover in month four that the data cannot be gathered. If your first idea is turned down, take the specific objection, apply the narrowing levers above, and return with a revised question rather than an unrelated new one.
Common mistakes when choosing a topic
The most frequent mistake is choosing a fashionable topic with no personal interest, which makes the long middle weeks of the project miserable. Close behind is choosing a question so safe it has been answered identically hundreds of times, leaving nothing to discuss. Students also over-promise methods — promising laboratory access, clinical data or senior interviewees they have not secured — and under-check the ethics timetable. A subtler error is copying a previous student’s topic almost unchanged; even where rules allow similar ground, you will be writing a project whose sources and findings are already settled, and originality marks will suffer. Treat your topic as a contract with yourself: interesting enough to sustain you, narrow enough to finish, and honest about the evidence you can truly reach.
Frequently asked questions
How specific should my dissertation topic be?
Specific enough that you can name the population, the setting and the data source in one sentence. If you cannot say who you would study or what evidence you would collect, the topic is still an area rather than a question and needs narrowing with the population, place, period or variable levers described above.
Can I change my topic after the proposal is approved?
Small refinements — adjusting a population or dropping one variable — are normal and usually need only your supervisor’s agreement. A complete change of subject normally requires a new proposal and fresh ethics approval, which is rarely practical after data collection has begun, so invest the checking time at proposal stage.
Is it a problem if my topic is similar to another student’s?
Similar areas are common and usually acceptable, especially on large courses, but identical questions in identical settings are not. Differentiate through population, setting, period or method, and make sure your literature review and data are genuinely your own.
What if I cannot think of a topic at all?
Return to the modules where your marks and interest were strongest, re-read your best assignment feedback, and list the debates or gaps those topics contained. Then test two or three draft questions against the feasibility checks. A structured starting point almost always produces options within a week of focused reading.
Does my topic need to be completely original?
No. At undergraduate and masters level, originality usually means a familiar question in a new context, with a new population, or tested with a different method. Chasing a question nobody has ever asked usually leads to a topic with no literature and no data — the two things your project cannot survive without.
Choose deliberately, narrow bravely and check feasibility before you commit, and your topic will carry you through the project instead of fighting you at every chapter. For guidance on what happens once the question is set — planning, chapters, methods and editing — see our full dissertation help page.