Ask three people how many references a dissertation needs and you will get three different answers: fifty, a hundred, or “anything under eighty will fail”. None is entirely wrong, and none is entirely right. A dissertation is not an essay stretched longer. It has its own structure and expectations, and its reference list has to do far more work. The honest answer scales with your level of study, your word count and your subject — and understanding why it scales is more useful than memorising any single figure.
This guide gives defensible ranges for undergraduate, master’s and doctoral work, explains where those references sit inside the dissertation, and shows what examiners actually judge when they turn to your bibliography. If you are still writing at essay length, the principles are the same but the numbers are smaller, as we explain in our guide to how many references an essay should have.
Why there is no single correct number
Universities rarely publish a required number, for good reason. A marker cannot judge scholarship by counting. What they can judge is whether you have engaged with the relevant literature thoroughly enough to justify your research question, choose a sound method and place your findings in context. Your reference list is evidence of that engagement. If the evidence is thin, the number looks wrong whatever it is. If it is strong, a list at the lower end of the usual range can still impress.
Ranges are still useful as a sense check. Think of density as roughly one source for every 250 to 350 words of sustained argument. That is not a university rule and it is not survey data; it is a practical model. In most academic writing, a new claim, comparison or piece of evidence arrives every few hundred words, and each moment usually needs support. Apply that to a long project with an extensive literature review and the total climbs quickly — which is why dissertation counts dwarf essay counts. Level matters too: an undergraduate dissertation shows you can work independently with scholarship, a master’s shows critical mastery of a field, and a doctorate must make an original contribution, which means knowing the field almost exhaustively.
Undergraduate dissertations: roughly 40 to 80
Most UK undergraduate dissertations run to 8,000 to 12,000 words, although 6,000-word and 15,000-word projects exist. For the typical project, roughly 40 to 80 sources is a defensible range. A shorter, tightly focused dissertation may sit comfortably at 40 to 50; a broader one, or one with a long literature review, may legitimately reach 70 or 80.
Fewer than 30 should give you pause. A very narrow, practice-based or primary-data project can sometimes manage with fewer, but in most cases a list that short means the literature review has summarised a handful of key texts instead of mapping a debate, identifying a gap and positioning your work within it. A list well over 100 prompts the opposite question: have you actually used all of these, or were some added to make the bibliography look impressive? Markers notice both problems.
Master’s dissertations: roughly 60 to 120
Master’s dissertations are commonly 12,000 to 20,000 words, with 15,000 a frequent requirement. Here, roughly 60 to 120 sources is a sensible range. A focused 12,000-word project might be strong with 60 to 80; a 20,000-word dissertation with an extensive literature review might need 100 or more.
The jump is not just quantity. At master’s level, examiners expect a higher proportion of peer-reviewed journal articles, a clearer grasp of competing theoretical positions, and evaluation rather than summary. Ten sources handled critically outweigh thirty summarised uncritically — but you still need breadth to show your judgements are informed. If the same eight or nine authors carry your whole literature review, that is a warning sign. Applied projects in education, business or health may also draw on policy and professional documents; those count if relevant and properly referenced, but they should supplement scholarly sources, not replace them.
PhD theses: 150 and well beyond
Doctoral theses operate on another scale. With limits often between 70,000 and 100,000 words and an expectation of comprehensive coverage, 150 sources is at the low end. Many theses cite 200, 300 or more. Humanities theses, engaging a long history of scholarship, can go higher; science theses may be moderated by a focus on recent literature, though citation density stays high. At this level the total is almost the wrong question. Examiners look for command: the seminal works, the current debates, the methodological literature and the newest publications. Missing an obvious key source matters more than finishing with 180 references instead of 210.
Where the references sit, chapter by chapter
A total can hide an unbalanced project, so consider the distribution. If you are still planning the project’s shape, check how long a dissertation should be chapter by chapter first.
Introduction. Few references are normal here. Cite some background literature to show why the topic matters, set out the problem and aims, and do not attempt to review the field.
Literature review. This carries the largest share — often around half your total, sometimes more. It must map what is known, show where scholars disagree, identify what has not been studied and build the case for your research. A 3,000-word review citing only twelve sources is almost certainly too thin, whatever your overall total says. Weave sources together thematically, compare methods and findings, and be critical throughout.
Methodology. Here you cite a different literature: research design, sampling, data collection, analysis, ethics and validity. Cite enough to show your choices were deliberate and grounded, including studies whose method you adapted.
Results and discussion. The results chapter cites selectively, because the focus is your own data. The discussion should reconnect findings to the literature: do they support, extend or contradict previous studies? A discussion with almost no references reads as descriptive rather than analytical.
Conclusion. Like the introduction, this needs few new references. Answer the question, summarise your contribution, acknowledge limitations and suggest future research, referring back to key sources rather than introducing new literature.
How subject changes the picture
Subject conventions explain much of the conflicting advice. In law, cases, statutes and other legal materials are authorities, often listed separately from secondary sources such as articles and textbooks. A law dissertation may have a modest secondary bibliography alongside a long table of cases and legislation — judging it on the bibliography alone would badly understate the research.
In laboratory sciences and other fast-moving fields, the weight falls on recent peer-reviewed articles. Books date quickly, so a textbook-heavy list looks out of step; expect a good proportion of sources from the last five to ten years, alongside foundational papers. In the humanities, books and chapters carry more weight and older scholarship may remain essential, while primary sources may form a separate list that a simple count does not capture. Social sciences, business, education and health usually mix articles, books and, where relevant, policy documents. Compare your list with good dissertations in your own subject, not a friend’s in another department.
What examiners actually judge
Examiners rarely count. They scan for signals. First, balance: seminal sources everyone would expect, plus current sources showing you are up to date. Only classics looks dated; only recent papers can suggest you do not know where the ideas came from. Second, relevance: every entry should be cited in the text and actually used, brought into the argument rather than mentioned in passing. Third, quality: peer-reviewed journals, respected publishers and authoritative official sources outweigh general websites and low-level textbooks. Textbooks have their place, especially at undergraduate level and for methodology, but a list built mainly on them will not convince anyone you have engaged with research.
They also notice accuracy. A recognised style applied consistently, with complete details, signals care; missing page numbers and untraceable sources signal the opposite. Cross-check that everything in the list is cited in the text, and everything cited appears in the list. Mismatches are common and easily avoided.
Padding, and the thin list behind a long word count
Padding means adding sources to hit a number in your head: listing books you never cited, dropping in famous but irrelevant names, or citing three sources for a point one could support. It is easy to spot — the citations feel decorative and the review reads like a catalogue, not a critical discussion. If a source did not influence your thinking or support a specific point, leave it out.
The opposite mistake is a thin list hiding behind a long word count: fluent writing built on fifteen or twenty sources recycled throughout. The prose may read smoothly, but long stretches of uncited assertion and a discussion that cannot compare findings with other studies give it away. Do not fix this by scattering citations through finished text. Read more widely and rebuild the sections that need support. A useful test is to pick any page at random and ask what it is based on. If the answer, page after page, is “my own opinion”, you need more sources. If it is a mix of evidence, theory and your analysis of both, you are in good shape.
Checking your own count
Count unique entries in the bibliography, not in-text citations — one source cited ten times is still one reference. Set the figure against your level, word count and subject, using the ranges here as guidance, not a verdict, and check the balance by chapter as well as the total.
For a quick sense check, use our free calculator to check a count against your word count and level of study. It cannot judge quality, but it will flag a list clearly out of line for a project of your size. If you are also writing shorter assignments, compare with our breakdown of how many references for a 3,000-word essay — the contrast shows why a dissertation cannot reuse essay habits. Remember what the number is for: references are the record of the conversation your dissertation joins. Aim for a list broad enough to show you understand that conversation, current enough to show where it is heading, and honest enough that every entry earned its place, and the count will look after itself.
Frequently asked questions
Is 50 references enough for an undergraduate dissertation? For most 8,000 to 12,000-word undergraduate dissertations, yes: 50 sits comfortably within the usual 40 to 80 range, provided the sources are relevant, well used and concentrated where they matter, especially in the literature review. Fifty strong sources beat eighty padded ones.
Can you have too many references in a dissertation? Yes, if the list includes sources never cited, only loosely relevant, or added purely to inflate the total. Examiners value selection and judgement, so a bloated bibliography can count against you just as a thin one can.
Do references count towards the dissertation word count? At most UK universities the bibliography is excluded, as appendices often are, but rules vary by institution and department. Check your assignment brief and handbook, and declare your word count as your university requires.
How many references should a 3,000-word literature review chapter have? There is no fixed rule, but the review carries the largest share of a dissertation’s sources, so only ten to fifteen is likely too narrow for most subjects. Many reviews of that length need substantially more to map the debate properly.
Should every reference in my list be cited in the text? In author-date styles such as Harvard, the list normally contains only works you cited. Some styles allow a bibliography of works consulted as well. Either way, nothing should be listed that played no part in the work, and everything cited must be listed in full.
How recent should dissertation references be? Aim for a mix: the seminal works your field is built on, however old, plus a good proportion of recent scholarship — in fast-moving subjects, much of it from the last five to ten years. Nothing recent looks out of date; no foundational sources looks uninformed.