The discussion is the chapter students most often misjudge in both directions. Some hand in a discussion barely longer than their conclusion, having reported their results twice and interpreted them never; others produce the longest chapter in the dissertation, swollen with material the reader has already met. Both errors cost marks for the same reason: the discussion is where a dissertation stops presenting work and starts arguing from it, and its length is a visible measure of how much arguing the writer actually did. This guide sets out the planning proportion, the counts it produces at the usual dissertation lengths, and the mechanics by which discussions grow fat or stay starved. To budget this chapter against the rest of the document, the Dissertation Chapter Word Count tool applies the same proportions to your total word count.
Length only is this guide’s territory. How to construct the chapter’s argument — the interpretation moves, the handling of limitations, the phrasing of implications — belongs to our guide on how to write a dissertation discussion chapter. Here the question is narrower: how much space that argument needs, and how to tell when yours has too much or too little.
The short answer: roughly a fifth to a quarter of the dissertation
Plan for the discussion to take roughly 20–25% of the total word count. Treat that figure the way every figure in this series should be treated: as a planning rule of thumb distilled from marker expectations, not a rule any university publishes. Your handbook, where it speaks, overrides it. But a discussion far below a fifth of the dissertation is almost always under-interpreting its findings, and one far above a quarter is almost always repeating something.
That share makes the discussion, alongside the literature review, one of the two largest chapters in the dissertation — and the pairing is not a coincidence. The literature review builds the argument the study will test: what the field knows, where it disagrees, where the gap sits. The discussion completes that argument: what your findings do to the field’s picture — confirm it, complicate it, or contradict it. Chapters that carry an argument need room to carry it in; chapters that report or describe can be shorter. The review takes 25–35% at the front of the dissertation, and the discussion takes 20–25% near the end, because between them they hold nearly all of the dissertation’s reasoning.
Word counts at the common dissertation lengths
10,000-word undergraduate dissertations: roughly 2,000–2,500 words. Twenty per cent of 10,000 is 2,000 and a quarter is 2,500. At undergraduate level this is usually the second-largest chapter, and supervisors who complain about a discussion almost always complain that this one is too short rather than too long.
12,000-word dissertations: roughly 2,400–3,000 words. The additional room over a 10,000-word project generally reflects more findings to interpret rather than a grander style: an extra research question, a second dataset, a subgroup comparison that each needs its pass through the literature.
15,000-word master’s dissertations: roughly 3,000–3,750 words. Master’s discussions are expected to do more theoretical work — engaging frameworks, not just individual studies — and that engagement is word-hungry. A 15,000-word dissertation with a 1,800-word discussion has usually reported a master’s-sized study and interpreted an undergraduate-sized one.
20,000-word master’s dissertations: roughly 4,000–5,000 words. At this scale the discussion often splits internally by research question or theme, each block with its own literature comparison, which is how the count is reached without padding. The chapter before it sets the scale for this one: our guide on how long a dissertation methodology chapter should be covers the 10–15% band the discussion’s findings were produced under.
What the words are actually spent on
A discussion is not one task but four, and they do not deserve equal shares. The opening recap of key findings should be brief — a few hundred words at most in a typical dissertation, enough to fix in the reader’s mind what is about to be interpreted, and no more. The interpretation itself is the chapter’s centre of gravity and should claim roughly half its length: finding by finding, what each result means, why it might have come out that way, and how it stands against the studies your literature review assembled. Implications for theory and for practice come next, shorter but not perfunctory. Limitations and the closing answer to the research question complete the chapter.
This internal balance is the real length test. A 2,500-word discussion that spends 1,500 of them recapping results and 400 on interpretation is, functionally, a 400-word discussion with an appendix of repetition. When students say their discussion “won’t reach” the expected length, they almost always mean the recap won’t; the interpretation, once genuinely attempted, has never yet failed to fill the space available for it.
Combined results and discussion chapters
Many shorter dissertations — and a great deal of scientific writing at every length — merge results and discussion into a single chapter, and the arithmetic changes accordingly. Instead of a results chapter of around 15% plus a discussion of 20–25%, the merged chapter typically carries roughly the sum of the two: somewhere in the region of 35–40% of the dissertation, organised as alternating blocks in which each result or group of results is reported and then immediately interpreted before the next is presented.
The merged format suits studies whose findings only make sense one at a time — a sequence of experiments, a set of assays, a series of model runs — because the reader needs each interpretation while the result is fresh. It punishes a different failure, though. In a merged chapter the reporting and the interpreting compete line by line, and weak writers let one devour the other: either the interpretation shrinks to a sentence per result because the next table is waiting, or the commentary swells until the results themselves become hard to find. If your department prescribes the combined format, budget the interpretation blocks explicitly — so many words per result — rather than trusting them to emerge.
Why discussions balloon
An overlong discussion is rarely overlong because its author had too much to say. It grows by repetition. The first source is results re-narrated at full length: means, percentages and quotations marched through a second time, as though the findings chapter were a rumour the discussion must confirm. The second is the literature review in miniature — studies summarised again from scratch instead of being invoked by name and finding, because the writer no longer trusts the reader to remember them. The third is implication inflation, where every finding sprouts recommendations for managers, policymakers and future researchers alike, each more speculative than the last.
The diagnostic is simple. Take any paragraph and ask whether it could have been written with the findings chapter closed. Interpretation cannot pass that test — it needs the results in view — but repetition passes it easily. Paragraphs that survive with the findings shut are the ones to cut, and cutting them usually returns a bloated chapter to its proper band without touching a single interpretive claim.
Why discussions starve
The underlong discussion fails quietly, and its shape is the mirror image. The findings are presented — again — and then the chapter stops, as if reporting were the same act as discussing. The tell-tale absence is the literature: studies from the review are never brought back, so no finding is ever declared consistent with, contrary to, or an extension of anything. Contradictions between the student’s own results and published ones go unremarked, which markers read not as modesty but as the student not noticing. Limitations appear as a list, unconnected to what they do to the conclusions. And the research question is answered only by implication, scattered across the chapter, never gathered into a verdict.
Starved discussions are commoner than bloated ones, and they share a cause: interpretation feels risky in a way reporting does not. To say what a result means is to commit to a judgement a marker can dispute; to repeat the result is safe. But the marks in this chapter live precisely in the committed judgements. A discussion at 12% of a dissertation is not a concise discussion. It is a chapter declining its job.
How discipline shifts the proportion
Laboratory sciences sit at or below the bottom of the band: their results arrive as figures and tables that carry weight prose would only duplicate, their conventions favour terse commentary tied closely to each experiment, and the combined format is common. Humanities and interpretive social sciences sit at or above the top, because there the discussion is not commentary on the argument; it is the argument — the place where readings, evidence and counter-readings are finally weighed. Health and applied fields occupy the middle, interpretation balanced against a substantial implications load, since findings there must speak to practice as well as theory.
Expanding a short discussion with analysis, not padding
If your discussion is under its budget, the remedy is more interpretation, and interpretation has reliable sources. Work the contradictions: where your findings depart from the strongest study in your review, spend words on why — population, method, context, timing. Entertain the alternative explanation for your headline result and argue, with evidence, why yours survives it; a discussion that has defeated a rival reading is longer and better than one that never faced one. Set boundary conditions: for whom, where and under what circumstances should your findings hold, and where should they not? Reason about mechanism — not merely that two variables moved together, but through what plausible process one produced the other. Every one of these moves adds words that earn marks, which is what separates expansion from the padding a marker can smell within a paragraph.
Discussion chapter too thin — or crowding out everything else?
Rebalancing a discussion is delicate work, because the chapter has to grow by adding judgement or shrink by cutting repetition, and neither fix is a matter of adding or deleting at random. If yours needs that surgery, United Academic Help offers dissertation help on individual chapters, with discussion chapters assessed against your findings and literature before any rewriting starts, from £20 per 1,000 words and the full pricing agreed first, before an order is placed. Write to team@unitedacademichelp.uk to place an order or send your chapter plan for a view on its proportions, or message us on WhatsApp via our contact page.
Frequently asked questions
How long should the discussion be in a 10,000-word dissertation? Roughly 2,000–2,500 words — a fifth to a quarter of the total. Around half of that should be interpretation against the literature; if the recap of findings is the longest part, the balance is wrong however respectable the total looks.
Should the discussion be longer than the results chapter? Usually yes. Results chapters report, and reporting is compact, especially where tables and figures carry the data; discussions interpret, and interpretation takes prose. A discussion shorter than its results chapter is a warning sign that findings have been presented but not examined.
How long is a combined results and discussion chapter? Roughly the two shares added together — in the region of 35–40% of the dissertation, so about 3,500–4,000 words in a 10,000-word project. Budget interpretation per result rather than as one block at the end, or the reporting will consume the space.
Can the discussion be longer than the literature review? Occasionally, in discussion-led or humanities-style dissertations where the later chapters carry the argument, but it is uncommon in empirical work. If yours is, check first that the review has not been squeezed below what it needs to set the discussion up — a long discussion built on a thin review tends to introduce literature late, which markers penalise.
Do limitations belong in the discussion or the conclusion? Both can hold them, at different depths. The discussion is where limitations are interpreted — what each one does to the confidence the findings deserve. The conclusion typically restates the headline limitations briefly as part of its closing account. Counting the full limitations treatment in the discussion is the safer plan.
My discussion is only 12% of my total. What is missing? Almost certainly the literature comparison. Go through your findings one by one and write, for each, where it agrees with the reviewed studies, where it departs, and why it might. That single exercise is where most under-length discussions find their missing words — as analysis, not filler.
Setting the discussion’s budget before the findings are in
Reserve the discussion’s share at the planning stage, before results exist to tempt you into endless reporting. A fifth to a quarter of the total, protected in advance, forces the useful question early: given roughly 2,500 words of interpretation in a 10,000-word dissertation, which findings genuinely earn a literature comparison and which can be answered in a paragraph? Run the whole plan through the Dissertation Chapter Word Count tool alongside the other chapters, and let the discussion’s budget discipline the findings chapter too — results that will never be discussed do not need three pages of presentation. A discussion planned this way arrives at its length by having enough to say, which is the only way of arriving there that markers reward.