Choosing a dissertation topic

Most advice on this starts with "choose something you are interested in", which is true and almost useless. Interest keeps you going in month four. It does not tell you whether the project is possible, and possibility is what actually decides your mark.

The four filters, in order

Run every candidate topic through these. In this sequence, because the cheapest rejections come first.

1. Can someone in your department supervise it?

Look at staff research profiles before you commit to anything. A topic nobody in the department works on means a supervisor who cannot tell you which literature matters, which methods are credible in that subfield, or that your question was settled in 2019.

This single filter removes more unworkable topics than any other, and it costs an hour.

2. Can you get the data or sources?

Ask the awkward question early: what exactly will I analyse, and how will I obtain it?

A topic requiring access you do not yet have is a topic with a deadline you do not control. Our guide to data collection methods sets out realistic ethics timelines.

3. Can it be answered in your word count and timescale?

Twelve thousand words sounds generous until you allocate it. Three thousand of those go to the literature review, and two thousand to method. You have perhaps five thousand words to say something. That is one question answered properly, not three surveyed.

4. Will you still care in month five?

Now interest matters. You will read about this topic for months, and a dissertation on something you find tedious is visibly worse by the discussion chapter.

Narrowing a topic until it works

Most first attempts are too broad. Narrowing happens along four dimensions.

The final version tells you the method, the participants and the analysis. The first version tells you nothing and would take a career. A question this precise is also what your research proposal will be built around.

The one-sentence test

State your question in a single sentence with no and. If you need and to join two questions, you have two dissertations.

The intervention question is a good dissertation. It is somebody else's. Note it in your "further research" section and move on.

Four topics that quietly fail

• The one with no literature. Feels original, usually means the question has been considered and dismissed, or the field calls it something else. Search harder, including in Google Scholar, before concluding you are first.
• The one with too much literature. Forty years of research on a well-worn question. You will spend the whole word count summarising and never reach a gap.
• The one that depends on one person saying yes. A single gatekeeper, employer or dataset owner. If they decline in week six you have no project.
• The one that is really a policy opinion. "Should the UK reform X" is an essay, not research, unless you have a method for answering it empirically.

Where to look for ideas

1. Your own module essays. Something you wrote 2,000 words on and wanted to keep going is a strong signal.
2. The "further research" sections of recent papers. Researchers state explicitly what needs doing next. It is the most underused source of dissertation topics there is.
3. Past dissertations in your department library. Not to copy, but to calibrate scope and see what your markers rewarded.
4. Staff research profiles. Working near a supervisor's interests means better supervision and faster answers.

Before you commit

Write the question down and check all five:

• Someone in the department can supervise it
• You know what you will analyse and how you will get it
• You can state it in one sentence without and
• You have found at least ten relevant recent sources
• You can describe the method in two sentences

If the last one is hard, the question is not yet specific enough. Go back to narrowing. An hour spent there saves a month in the middle of the project.

Last updated: 30 July 2026 · Written by the AssignWise editorial team ·Spotted something wrong?