The conclusion is the last thing your reader sees, and often the part they remember best. After pages of methods, results, and discussion, a tired reader turns to the conclusion to answer one simple question: so what did this study actually show, and why does it matter? A strong conclusion answers that question in a few clear paragraphs, without repeating the whole paper and without claiming more than the evidence supports.
This guide explains how to write that kind of conclusion. You will learn what a conclusion must do, how to restate your contribution, how to handle limitations honestly, and how to end with implications that readers and reviewers can trust. Each section ends with a practical point you can apply to your own draft today.
A conclusion has one job: to close the argument your paper opened. It should remind the reader of the question you asked, state the answer your evidence supports, and explain why that answer matters. It is a synthesis, not a summary — it pulls the findings together into a single, confident statement of what the study contributes, rather than listing each result again in the order it appeared.
Just as important is what a conclusion must not do. It must not introduce new data, new citations, or new arguments that were never discussed in the body of the paper. It must not repeat the abstract word for word, and it must not end with vague phrases such as "more research is needed" without saying what research, and why. A conclusion that adds surprises at the last minute confuses readers and gives reviewers an easy reason to ask for revision.
Many authors open their conclusion by copying sentences from the abstract. Readers notice, and the paper feels padded rather than finished. Instead, restate your contribution in fresh words, and in a form that reflects everything the paper has now shown. The abstract promised an answer; the conclusion delivers it, with the weight of the full argument behind it.
To do this well, write one sentence that states what is now known that was not known before your study, and one sentence that says how you established it. Keep the language specific: name the population, the setting, or the method where it matters, and avoid grand claims. A contribution stated precisely is far more persuasive than one stated broadly, because a precise claim is one a reader can check against your results.
Most papers report several findings, and the conclusion is where they must come together. Synthesis means showing how the individual results combine into one coherent answer to your research question. Ask yourself: if a reader remembered only one sentence from my paper, which sentence should it be? That sentence belongs at the heart of your conclusion, and every other sentence should support it.
Where findings point in different directions, say so plainly rather than hiding the tension. A result that partly supports your hypothesis and partly contradicts it is still a finding, and explaining the pattern honestly strengthens your paper. Readers trust authors who report complexity fairly; they distrust conclusions that smooth over inconvenient results in order to sound tidier than the data allow.
Your introduction identified a gap in the literature — something unknown, uncertain, or unresolved that your study set out to address. The conclusion should close that loop by showing, explicitly, how the gap has now been narrowed. This bookend structure is what makes a paper feel complete: the reader can see that the promise made at the start has been kept at the end.
Be honest about the size of the step. Most studies narrow a gap rather than closing it entirely, and saying so is a strength, not a weakness. Explain what part of the gap your study addressed, what remains open, and why the remaining questions matter. A conclusion that claims to have solved a whole field's problem from a single study will not convince a specialist reader, while one that names its exact contribution will.
Every study has limitations — a small sample, a single setting, a measurement that was imperfect, or a design that cannot prove cause and effect. Stating them in the conclusion, briefly, shows that you understand your own evidence and its boundaries. Reviewers already look for limitations; naming them yourself demonstrates judgement and saves the reviewer from having to point out what you already know.
Keep limitations short and specific, and explain how each one affects what can be claimed. Avoid apologising at length or listing every minor weakness, which can undermine confidence in the work. One or two sentences are usually enough: name the limitation, state its effect on the interpretation, and move on. The goal is a balanced conclusion, not a confession, and a paper that is honest about its limits is easier to trust.
Implications answer the reader's next question: what should change because of this study? For practitioners, that might mean a change in how a procedure is carried out, a diagnostic step worth adding, or a practice worth reconsidering. For policymakers, it might mean evidence that supports a guideline, a training need, or a direction for future funding. Name the audience for each implication so readers can see who should act.
Keep implications proportionate to your evidence. A single small study rarely justifies a sweeping policy change, and overstating implications is one of the fastest ways to lose a reviewer's confidence. Use careful language — "suggests", "supports", and "indicates" rather than "proves" — and tie each implication directly to a finding reported in your results. An implication that cannot be traced back to your data does not belong in the conclusion.
Future research suggestions should grow naturally out of your study's limitations and unanswered questions. The most useful suggestions are specific: the population that should be studied next, the variable that should be measured differently, or the design that could test the mechanism your study could only hint at. Specific suggestions help other researchers and show that you understand where your field should go next.
Avoid the empty formula of calling for "further research" with no direction, and avoid promising work you do not intend to do. Suggestions are not commitments, and reviewers read them as a sign of scholarly honesty rather than as a plan. Two or three well-chosen directions are enough; a long wish list dilutes the conclusion and makes the paper feel unfinished rather than forward-looking.
A conclusion should be short relative to the rest of the paper. For most journal articles, a few paragraphs are enough — often around five to ten percent of the total length. A conclusion that runs for pages suggests the author has not yet decided what the paper's main point is, while a conclusion of a single vague sentence suggests the argument was never finished. Aim for the middle: long enough to state the answer and its implications, short enough to be read in one sitting.
Length also signals confidence. When you can state your contribution in a few clear sentences, readers believe you understand your own work. Before submitting, try cutting your conclusion by a third: remove repeated results, generic background, and hedging that adds no meaning. If the shortened version still says everything important, keep it short. Brevity is not laziness; it is the final act of editing.
The most common reason conclusions fail in peer review is a mismatch with the results. Every claim in the conclusion must be traceable to a result reported earlier in the paper, with the same strength and the same scope. If your results show an association, your conclusion must not claim a cause. If your sample came from one region, your conclusion must not speak for a whole country. Reviewers check this alignment carefully.
To test your own draft, go through the conclusion sentence by sentence and ask: where in the results does this appear? If a sentence has no matching result, either delete it or soften it until it matches. Pay special attention to numbers, effect sizes, and qualifiers — a conclusion that repeats the numbers from the results accurately is far more convincing than one that rounds them into a stronger-sounding claim.
Several mistakes appear again and again in submitted conclusions. The first is introducing new evidence or citations at the very end, which forces the reader to reopen an argument that should be closed. The second is repeating the abstract or the discussion almost word for word, which wastes the reader's time. The third is overstating the findings, turning a modest association into a breakthrough, or a single study into a final answer.
Other mistakes are subtler but just as damaging: ending with an apology for the study's weaknesses, adding a long list of future work that reads like an admission of incompleteness, or closing with a quotation or anecdote that belongs in an essay rather than a research paper. As a warning sign, treat any service or message that promises a guaranteed acceptance as a red flag — no honest service can promise an editorial outcome. Keep the ending factual, confident, and grounded in your own results.
Consider a weak conclusion for a hypothetical study of feeding practices on small dairy farms: "This study shows that better feeding improves milk yield. Farmers should change their feeding practices. More research is needed in this area." It is vague, it overstates the finding, it gives no detail, and its final sentence says nothing useful. A reader learns almost nothing about what the study actually found.
A stronger version of the same conclusion would read: "Among the surveyed farms, herds receiving the supplemented ration produced higher average daily milk yield than herds on the traditional ration, after accounting for herd size. This suggests that ration supplementation is a practical option for smallholders in similar settings, though the finding comes from one season and one region and should be tested across seasons before wider recommendations are made. Future work should examine cost-effectiveness, which this study did not measure." Notice the difference: the claim is specific, the limits are named, and the next step is concrete.
Before you submit, read your conclusion on its own, as a reviewer would, and check it against a simple standard. Does it state the answer to your research question in one clear sentence? Does every claim match a result in the paper? Are limitations stated briefly and honestly? Are implications tied to the evidence and addressed to a clear audience? If any answer is no, revise before submitting rather than hoping a reviewer will not notice.
Also check the mechanics: no new citations, no new data, no numbers that differ from the results, and no claim stronger than the design allows. Read the conclusion aloud — a sentence you stumble over will also trip your reader. Finally, ask a colleague to read only the introduction and the conclusion together. If they can tell you what the paper set out to do and what it found, your conclusion is doing its job.
| Element | What to check |
|---|---|
| Answer | States the answer to the research question clearly |
| Contribution | Restates what is new, in fresh words |
| Fit | Every claim matches a reported result |
| Limits | Limitations stated briefly and honestly |
| Implications | Practical and policy implications tied to evidence |
| Next steps | Specific future research directions, not vague calls |
| Length | Short, proportionate, and free of new material |
These authoritative external resources explain the standards and tools mentioned in this guide:
How long should a research paper conclusion be?
For most journal articles, a few paragraphs are enough — often around five to ten percent of the total length. It should be long enough to state the answer, the contribution, the limitations, and the implications, and short enough to be read in one sitting.
Should a conclusion repeat the abstract?
No. The abstract promises an answer; the conclusion delivers it in fresh words, with the weight of the full argument behind it. Repeating the abstract word for word makes the paper feel padded rather than finished.
Can I include new citations in the conclusion?
As a rule, no. The conclusion should not introduce new evidence, new citations, or new arguments that were never discussed in the body of the paper. Anything important enough to cite belongs in the introduction or discussion.
Should I mention limitations in the conclusion?
Yes, briefly. Naming one or two key limitations and explaining how they affect interpretation shows judgement and builds trust. Keep it short — a balanced conclusion is not a confession.
How do I end a conclusion strongly?
End with the implication that matters most: what should researchers, practitioners, or policymakers take from your study, stated in proportion to your evidence. Avoid vague closing lines such as "more research is needed" without saying what research and why.
What if my results are mixed or inconclusive?
Report the pattern honestly and explain it. A conclusion that acknowledges complexity and names the next step is stronger than one that smooths over inconvenient results. Mixed findings still contribute, because they narrow the question for future work.