Choosing a dissertation topic you will not regret

The dissertation topic is the highest-leverage decision of your degree, and it is usually made in a fortnight, under pressure, on the basis of what sounded interesting in a lecture. A good choice makes the following nine months hard but tractable. A bad one makes them a slow argument with a question that cannot be answered.

Choosing well is not about finding the most original idea. It is about finding the intersection of four things: something you can stay interested in, something you can actually do, something with enough evidence available, and something someone in your department can supervise. This guide shows you how to find that intersection and how to narrow a topic into an answerable question.

Choose a question you can survive for nine months

Dissertations are endurance events. Around month five, in the middle of transcription or a chapter that will not cohere, the only thing keeping you at the desk is some residual curiosity about the answer. Choose a topic that bores you slightly now and it will be intolerable then.

The reliable test is whether you find yourself reading about it when you do not have to. If you have ever gone down a rabbit hole about local housing policy, sports analytics, the history of a particular medical treatment or how a company failed, that is real evidence about your attention — worth more than a list of topics that sound impressive.

That said, interest alone is a trap. Plenty of fascinating questions cannot be answered by one student in nine months with no budget. Interest is the first filter, not the only one.

Feasibility beats ambition, every time

The most common cause of a distressed dissertation is a design that depends on something outside the student's control: an organisation granting access, a hard-to-reach population agreeing to be interviewed, a dataset that turns out to cost money, or ethics approval on a timescale that does not fit the calendar.

Before committing, work out what you would need and whether you can get it. If your study requires forty participants from a specialised profession, how will you reach them, and what happens if only nine reply? If it requires company data, who signs that off, and have you asked? A design that can be executed alone with publicly available material is not a lesser dissertation; it is frequently a better one, because the time goes into analysis rather than logistics.

Secondary-data and documentary projects are consistently under-chosen and consistently score well. Government datasets, published corpora, court judgments, company annual reports, parliamentary records and existing survey series are all rich, free, immediately available and unaffected by whether anyone replies to your emails.

• Can I obtain the data alone, without anyone's permission? If not, who must say yes, and by when?
• What ethics route does this need, and how long does that route take at my institution?
• What does this cost — in money, travel and software?
• If my main data source falls through in week six, what is my plan B?

The four-filter test

Run every candidate topic through four filters in order, and drop anything that fails one. Interest: will I still care in March? Feasibility: can I complete this alone, on time, with what I have? Evidence: does enough material exist to build an argument, and can I reach it? Supervision: is there someone in the department who knows this area well enough to help?

The supervision filter is the one students most often ignore, and it matters more than the topic's inherent quality. A supervisor working near your area will spot the flaw in your design in ten minutes, point you at the three papers you have missed, and know which examiners are likely to read it. A supervisor who is simply available will give you administrative support and little else.

Look at your department's staff pages before you finalise the topic, not after. Read two or three abstracts from anyone whose work is nearby. Choosing a topic that sits next to an existing supervisor's expertise is not a compromise — it is the single most effective thing you can do to raise your final mark.

Narrowing: from topic to answerable question

'Sustainability in fashion' is a topic. It is not a question, and it cannot be answered. Narrowing means adding constraints along four dimensions until what remains is answerable: population or case, concept, outcome, and time or place.

Watch the narrowing happen. Sustainability in fashion becomes sustainability claims in fashion marketing; becomes how UK fast-fashion retailers frame environmental claims in advertising; becomes how three UK fast-fashion retailers' environmental claims in Instagram advertising changed following the CMA's Green Claims Code, 2021 to 2026. The final version tells you what to collect, how much of it, and when to stop.

New researchers almost always overestimate scope. If your question could plausibly describe a three-year funded project, cut it again. A narrow question answered thoroughly is a first; a broad question answered thinly is a 2:2, regardless of how much work went into it.

Pressure-test it before you commit

Write a one-page summary — question, why it matters, how you would answer it, what data you need, what could go wrong — and take it to three different people: your prospective supervisor, a friend on your course, and someone who knows nothing about the subject.

Each will catch something different. The supervisor spots the flaw in the design and the literature you have missed. The coursemate tells you whether the scope resembles other dissertations on your programme. The outsider tells you whether the question makes sense to a human being, which is a surprisingly good proxy for whether it is well formed.

Come to the supervisor meeting with two or three candidate questions rather than one. Presenting alternatives changes the conversation from approval-seeking to genuine advice, and supervisors give better advice when they are comparing options than when they are being asked to bless a decision you have visibly already made.

When to change topic, and when to hold your nerve

Almost every dissertation goes through a period around weeks six to ten when the topic feels wrong. Usually it is not wrong; you have simply learned enough to see the difficulties, which is a sign of progress rather than of a bad choice. Wanting the topic to be more interesting is not a reason to change.

There are real reasons to change, and they share a common feature: the project cannot be completed as designed. Access has definitively fallen through with no alternative. Ethics approval has been refused. The question has already been answered comprehensively by a study you have just found. The data does not exist. Those are structural problems, and the earlier you act the cheaper they are.

In most of those cases the fix is a pivot rather than a fresh start. Keep the field, keep the literature you have read, and change the method or the case: interviews become documentary analysis, one organisation becomes a comparison of published reports, a national question becomes a regional one. A pivot in week eight costs a fortnight; a new topic in week eight costs the dissertation.