The results are in, the tables are built, and now comes the section students dread most: the discussion — where you stop reporting what you found and start saying what it means. That shift, from description to interpretation, is exactly why so many discussion sections go wrong. Some students repeat their results in different words. Others overclaim, treating a small study as though it settled the field. And some bury the interesting finding under a page of hedging.
A good discussion does five things: it restates the main findings, explains what they mean, compares them with the existing literature, acknowledges the study’s limitations honestly, and points to what should happen next. This guide shows you how to do each one.
What a Discussion Section Actually Does
Think of the discussion as the section where your paper earns its keep. The introduction promised an answer to a research question; the methods showed how you looked for it; the results showed what you found. The discussion delivers on that promise — you interpret the findings, place them in the context of what is known, and say why they matter.
That is also why the discussion is usually the most referenced section after the literature review. Nearly every interpretive claim should be anchored to a source: a study that agrees with you, one that contradicts you, or a theory that explains your finding. A discussion with no citations is just opinion; one that cites only agreeing studies looks like cherry-picking. The University of Southern California’s guide to writing the discussion section puts it plainly: discuss the significance of your findings in the same sequence you described them in the results.
The Inverted Pyramid Structure
The most reliable shape for a discussion is the inverted pyramid: start with your specific findings, then widen out to the literature, then to theory, then to practice. Here is the sequence.
Step 1: Restate the research problem and answer it
Open by briefly restating your research question and giving the direct answer your results support — one paragraph, declarative, no hedging: “This study asked whether daily Instagram use predicts body-image dissatisfaction among first-year female undergraduates. The results indicate that it does — but the effect is driven by appearance-focused engagement, not by total time online.” The reader should know your headline finding within the first five lines. If a question remains unresolved because the data could not answer it, say so here, and say why.
Step 2: Explain what the findings mean, one by one
Work through your major findings in the same order you reported them in the results. For each one, state the finding, then interpret it: what does this result actually tell us, and why does it matter? A finding on its own is a fact; explaining it is your contribution.
Start with your most significant or unexpected finding if it deserves the emphasis — the one exception to following the results order. Give each finding at least two or three sentences of interpretation; one sentence per finding is a summary, not a discussion.
Step 3: Compare with the existing literature
For each finding, ask how it relates to what previous studies found. Does it confirm earlier work? Contradict it? Extend it to a new population? Revisit the key sources from your literature review here — and bring in studies you saved for the discussion because they matter more as comparisons than as background.
Agreement is easy to write; disagreement is where the marks are. If your results differ from a published study, do not ignore the difference. Explain it: different sample, different method, different context, or a genuine challenge to the earlier conclusion. A contradiction you explain is a strength. A contradiction you hide is a problem your examiner will find.
Step 4: Acknowledge limitations
Every study has limitations, and naming them yourself is far better than having them pointed out. Cover the ones that genuinely affect your interpretation: a small sample, a single-site design, self-reported data, a short time frame. Be specific — “the sample was drawn from a single university” is a real limitation; “more research is needed” is not. Keep the tone honest, not apologetic: you are describing the boundaries of what your study can claim.
Step 5: State the implications and suggest further research
Close by stating what your findings mean in practice and what should happen next. Implications might be practical — a recommendation for teaching, policy, or clinical practice — or theoretical, such as a refinement to an existing model. Then suggest specific further research: a larger sample, a different population, a longitudinal design, a variable your study could not measure. Link each suggestion to a limitation from your discussion, so it reads as a natural next step.
A Worked Example: Discussion Paragraphs in Action
Take a study asking whether daily Instagram use predicts body-image dissatisfaction among first-year female undergraduates. Here is how its discussion might open.
Restate and answer: “This study asked whether daily Instagram use predicts body-image dissatisfaction among first-year female undergraduates at UK universities. The results indicate that it does: students reporting more than two hours of daily use scored significantly higher on body dissatisfaction than lighter users. But the effect was driven almost entirely by appearance-focused engagement — viewing influencer content and editing photos — not by total time online.”
Interpret: “This distinction matters. It suggests the mechanism is social comparison with idealised images, not screen time as such: a student who spends two hours messaging friends shows no comparable effect, while one who spends thirty minutes browsing influencer feeds does. That finding refines the common claim that ‘social media harms body image’ into something more precise — and more actionable.”
Compare: “The result aligns with Fardouly et al. (2017), who found that appearance-focused social media use predicted body dissatisfaction in young women, but extends their work to first-year students — a period when identity and peer comparison are in flux. It differs from earlier surveys that treated total usage time as the key variable, likely because those studies never separated types of engagement.”
Handling Unexpected Results
An unexpected result is not a failed study — it is often the most interesting part of the paper. Handle it in three moves. First, describe it plainly: “Contrary to the hypothesis, students who reported the heaviest social media use did not show the highest dissatisfaction.” Second, suggest why it might have appeared — a measurement issue, a subgroup effect, an unanticipated variable. Third, state its significance: does it change your answer to the research question, or qualify it?
Six Discussion Mistakes (and How to Fix Them)
Restating the results. Repeating numbers in prose. Fix: mention each finding once, in a “bridge” sentence, then spend the paragraph on what it means.
Introducing new results. Presenting a finding for the first time here. Fix: anything you discuss must already exist in the results section.
Overclaiming. “This study proves that…” from forty students at one university. Fix: use “suggest” and “indicate”, and state the scope of your sample.
Ignoring contradictory literature. Citing only studies that agree with you. Fix: for each major finding, find one study that complicates it and explain the difference.
Hiding the limitations. No limitations paragraph. Fix: write one — its absence reads as carelessness or overconfidence.
Repeating the conclusion. Ending with a full summary. Fix: end with implications and further research — the summary belongs in the conclusion, as the UNC Writing Center’s guide to writing conclusions explains.
Discussion vs Conclusion: Where Does One End?
Where does the discussion stop and the conclusion begin? The difference is one of scope and job. The discussion interprets: it explains findings, compares them with the literature, and weighs limitations in detail. The conclusion synthesises: it gives the short answer to the research question, states the study’s contribution in a few sentences, and stops — no new interpretation, no new citations, no new limitations.
A practical test: if a paragraph explains why a finding matters or how it compares with earlier work, it belongs in the discussion. If it states what the study found and why that matters in one or two sentences, it belongs in the conclusion. When the discussion is written well, the conclusion almost writes itself. Our guides to writing a research paper abstract and writing a research paper introduction cover the other ends of the paper.
A Pre-Submission Checklist
Before you submit, run this check on your discussion section. Does it open with a direct answer to the research question? Is every finding interpreted, not just restated? Are the findings compared with previous studies, including ones that disagree? Are limitations stated specifically, with their consequences? Does it end with concrete implications and specific suggestions for further research? And does it contain no new results and no claims the evidence cannot support? Six yeses means the discussion is doing its job.
The discussion is where a research paper stops being a report and becomes an argument. Write it as one: findings first, interpretation second, literature third, limitations honestly stated, implications clearly drawn. Do that, and the section students dread becomes the section that earns the most marks.
Struggling to interpret your results or turn them into a discussion that holds together? The Pen Master pairs you with a subject-matched academic writer who can help you write the discussion, conclusion, and abstract of your research paper, with free revisions until it reads the way you want. Message us on WhatsApp at +49 1521 3962850 for a quick quote, or send your brief through the contact form — we will get back to you the same day.