For a 15,000-word dissertation, a sensible working range is roughly 45 to 60 references, with many strong dissertations landing between 50 and 70 depending on the subject. That range is not a rule set by any university, and no marker counts your references and awards marks for hitting a number. It is a planning figure, derived from the density of sourcing that a dissertation of this length normally needs, and it is most useful at the start of your research, when you are deciding how much reading to do and how to spread it across your chapters.
The reason the range is wide is that different subjects cite at very different rates. A literature-heavy dissertation in history, sociology or English may cite a source every 200 words or so, because almost every paragraph builds on someone else’s argument. An empirical dissertation in psychology or health sciences may cite less densely in its results chapter, where the data speaks for itself, and more densely in its literature review and discussion. A practice-based or design dissertation may cite fewer sources overall and lean harder on primary material of its own. Your supervisor’s expectations and your department’s past dissertations matter more than any general figure, so treat the range below as a starting point and calibrate it against examples from your own course.
Where the 45 to 60 figure comes from
The usual rule of thumb behind reference calculators is one reference for every 250 to 350 words of main text. At one reference per 300 words, a 15,000-word dissertation needs 50 references. At the denser end, one per 250 words, it needs 60. At the lighter end, one per 350 words, it needs about 43. That is how the 45 to 60 range is produced, and it is also why a single exact number would be misleading: the right density depends on how much of your word count is literature, analysis and discussion rather than method, results or description.
It helps to compare with shorter pieces you have already written. A 3,000-word essay with 12 to 15 references sits at roughly one reference per 200 to 250 words, which is denser than a dissertation, because nearly every paragraph of a short essay is making an argument from the literature. A dissertation spreads its word count across chapters that do different jobs, so its overall density is lower even though its literature review chapter may be just as dense as an essay. If you applied essay density to a whole dissertation, you would end up with 75 or more references, which is possible in some subjects but is not a target to chase for its own sake.
Published guidance that circulates online often quotes higher figures, including a claim that a 15,000-word dissertation needs around 95 references. That figure comes from applying a very dense rate of roughly one reference per 160 words to the entire document, which is far denser than most markers expect outside a systematic literature review. Other sources suggest 50 to 100 references for a master’s dissertation of 15,000 to 25,000 words, which is consistent with the range in this guide at the 15,000-word end. The honest position is that there is no official number, and a dissertation with 48 well-used references will always beat one with 95 references that are name-dropped and never discussed.
Spreading references across your chapters
A 15,000-word dissertation is usually divided into an introduction, a literature review, a methodology, results or findings, a discussion and a conclusion, though the exact structure varies by subject. References do not spread evenly across those chapters, and planning the spread is more useful than planning a single total.
The literature review carries the heaviest load. At 3,000 to 4,500 words, it might hold 20 to 30 references, because its whole job is to map what is already known, identify disagreements and justify your research question. The introduction, at 1,000 to 1,500 words, might hold 5 to 10 references that set the context and point to the gap you are filling. The methodology chapter, at 1,500 to 2,500 words, typically holds 8 to 12 references, mostly to methods texts and to studies that used a similar design. Results or findings chapters cite the least, often only a handful of references, because they report your own data. The discussion then returns to the literature, with 10 to 15 references that compare your findings with earlier work. The conclusion rarely adds new references at all.
Add those chapter figures together and you land in the 45 to 60 range, which is a useful cross-check on the per-word calculation. If your plan shows 70 references but 50 of them sit in the literature review, your review is probably trying to cover too much ground and your discussion will have nothing left to say. If your plan shows 30 references and your literature review holds only 10, you probably have not read widely enough to justify your question. The chapter spread exposes problems that a single total hides.
What changes the number: subject, level and method
Level matters. A 15,000-word undergraduate dissertation and a 15,000-word master’s dissertation are not judged by the same standard, even at the same length. Master’s work is expected to show greater critical engagement with the literature, which usually means a slightly higher reference count and, more importantly, deeper discussion of each source. An undergraduate dissertation at the lower end of the range, around 45 to 55 references, can be excellent if every source is genuinely used. A master’s dissertation at the same length would more often sit at 55 to 70.
Method matters too. A systematic review or a dissertation built entirely from secondary sources will cite far more than an empirical study, because the literature is the data. A qualitative interview study with a small sample may cite fewer sources but discuss each one at greater length. A quantitative study with a large dataset may cite a moderate number, concentrated in the literature review and discussion, with the analysis chapters standing on their own. None of these patterns is a weakness. The mistake is to force an empirical dissertation up to a review-style reference count by padding the list, or to starve a literature-based dissertation of sources because you stopped reading once you reached an arbitrary number.
Subject conventions matter as well. In law, history and literary studies, close engagement with primary texts and commentary can push the count higher, and footnotes may carry additional references that do not appear in the same way as in-text citations. In engineering, computer science and some applied sciences, a dissertation may cite fewer, more technical sources, including standards, manuals and datasets, and the quality of those sources matters more than their number. If your department publishes past dissertations, reading two or three of them and counting their references will tell you more about your local norm than any general guide.
Quality beats quantity, and markers can tell
The most common referencing mistake in long dissertations is not having too few references but having too many shallow ones. A reference that appears once, in a list of five citations at the end of a sentence, with no discussion of what that source actually argued, adds almost nothing. A reference that is introduced, explained and then returned to in the discussion does real work. Markers notice the difference quickly, and they notice padding even faster: a reference list full of sources that never appear in the text, or a paragraph that cites ten sources for a claim that needs two, reads as insecurity rather than scholarship.
A practical test for every reference on your list is to ask three questions. Did you actually read it, or only its abstract? Does it support a specific point in your argument, and could you say which point? Would removing it weaken your dissertation? If the answer to any of those is no, the reference is a candidate for cutting, however impressive it looks on the list. Cutting weak references is not losing marks; it is tightening the argument, and it often improves the grade.
There is also a difference between references and sources consulted. Your reference list should contain only works you cited in the text, unless your department explicitly asks for a separate bibliography of background reading. Keeping a private reading log of everything you consulted is excellent practice, but that log is not your reference list, and merging the two inflates your count with works you never used.
Planning your reading to hit the range honestly
A workable approach is to plan backwards from your chapters. Sketch your chapter word counts first, then assign each chapter a reference budget using the spread above, and then read to fill those budgets with sources you actually need. Start with the literature review, where the budget is largest, and read until you can map the main debates, the key disagreements and the gap your study addresses. Stop adding sources to a chapter when new reading stops changing your argument, which is the point at which you have read enough for that chapter, whatever the number happens to be.
Keep a running count as you write, not at the end. Most reference managers show how many entries you have cited, and a quick check at the end of each chapter tells you whether a chapter is running thin or heavy. If a chapter is running heavy, the usual cause is that you are summarising sources instead of synthesising them: grouping two or three sources that make the same point into one sentence, and then discussing what they agree and disagree about, cuts the count without losing content. If a chapter is running thin, the usual cause is that you are asserting things without evidence, and the fix is more reading, not more citations to the same few sources.
Before you submit, do a final audit. Check that every in-text citation has a matching reference list entry and vice versa, check that you have not cited the same source under two different spellings or years, and check that your reference list is alphabetical by surname. A clean list of 50 references looks far more professional than a messy list of 70, and it signals to your marker that the same care went into the rest of the dissertation. To plan the numbers for your own word count and chapter structure, use the Reference Count Calculator to work out a target range before you start writing.
Frequently asked questions
Is 50 references enough for a 15,000-word dissertation? Yes, for most subjects. Fifty references at 15,000 words is one per 300 words, which sits in the middle of the normal range. What matters is whether those fifty are well chosen, genuinely read and properly discussed, not whether you could have added twenty more.
Can I have too many references? Yes. A list far above 80 or 90 for a 15,000-word dissertation usually means sources are being name-dropped rather than discussed, or that background reading has been merged into the reference list. If your count is very high, check each entry against the three-question test in this guide and cut the ones doing no work.
Does the reference list count towards the 15,000 words? Usually the reference list itself is excluded from the word count, while in-text citations are included. Check your own assignment brief, because some departments count citations differently, and a small number of courses include the list. Do not assume either way.
Do references in the literature review count differently? No, there is no separate counting rule, but the literature review naturally carries more references than other chapters because mapping existing work is its purpose. A literature review with very few references is a warning sign that the reading has not been done, whatever the total for the dissertation.
What if my supervisor asks for a specific number? Follow your supervisor. A departmental expectation, even an informal one, outweighs any general guide, because it reflects what your markers will actually look for. If the number they suggest seems very high or very low for your subject, ask them what kind of sourcing they have in mind before you change your plan.
Should I count sources I read but did not cite? No. Your reference list should contain only works cited in your text, unless your department asks for a separate bibliography. Keep background reading in a private log, and only promote a source to the reference list when you actually cite it in your argument.