How To Write a Scientific Discussion for a Research Paper

1. The Role of the Discussion Section

In the methods and results sections of a scientific paper, your job is largely descriptive: you explain how the study was carried out and what you found. The discussion section is different. Here you are expected to interpret your findings, explain their meaning, evaluate their significance, and show how they relate to existing knowledge. A well-written discussion demonstrates not only that you have obtained data, but that you can think critically about them.

A strong discussion:

• revisits your central research questions and hypotheses,
• summarises and interprets the most important results,
• situates your findings within the published literature,
• explores implications for theory, practice, and future research,
• acknowledges limitations honestly, and
• ends with clear, concise take-home messages.

Finding the right balance is crucial: too short, and your discussion will seem superficial; too long, and it may feel repetitive or speculative. The “dos and don’ts” below are designed to help you achieve that balance.

2. Restating and Framing Your Results

Do: Restate key findings briefly and interpret them

It is useful to begin the discussion with a succinct overview of your most important results, especially those that are unexpected or directly address the main research question. However, the discussion is not a second results section. When you restate findings, immediately shift into interpretation.

For example, instead of repeating “Group A had a mean score of 15.3 and Group B had a mean score of 20.7,” you might write:

“Participants who received the intervention performed substantially better on the memory test than those in the control group, suggesting that the training protocol had a measurable impact on short-term recall.”

Here you still convey the result, but you also point towards its meaning.

Do not: Copy the results section word for word

A common mistake is to copy sentences directly from the results section into the discussion, sometimes with only minor changes. This produces tedious repetition and does not provide the interpretation readers expect. The discussion should move beyond numbers and p-values, focusing on “what these findings tell us” rather than “what exactly we measured.”

3. Linking Back to the Introduction

Do: Revisit research questions, problems, and hypotheses

Your discussion should clearly answer the questions raised in your introduction. If you started by asking whether a particular treatment improves outcomes, your discussion should explicitly state whether your data support that expectation, and to what extent. You can also revisit theoretical frameworks or debates introduced earlier and explain how your results support, refine, or challenge them.

For instance:

“This study set out to test the hypothesis that X. Our findings support this hypothesis in younger adults, but not in older participants, indicating that age may moderate the effect.”

Do not: Repeat the introduction or introduce new problems

You do not need to retell the full background story. Avoid repeating long passages from the introduction or adding entirely new research questions that were not previously mentioned. If your interpretations lead you towards new concepts or directions, consider revising the introduction so that it foreshadows these themes instead of letting them appear suddenly in the discussion without context.

4. Interpreting Your Results

Do: Explain what the results mean and why they matter

The heart of the discussion is your interpretation. Make it clear:

• how your results answer your research questions,
• why the findings are important for theory, practice, or policy, and
• what patterns, relationships, or surprises they reveal.

Unexpected findings often deserve special attention. Rather than apologising for them, explore why they might have occurred and what they suggest about the phenomenon under study. Considering alternative explanations also strengthens your argument, showing that you have weighed competing interpretations.

Do not: Over-interpret or force unsupported conclusions

There is a fine line between creative interpretation and wishful thinking. Avoid claiming effects that your data do not support, or stretching modest findings into sweeping generalisations. For example, a small, single-site study should not be presented as definitive proof that a treatment will work in all settings. Overstating the importance of your results can undermine your credibility and draw criticism from reviewers and readers.

Similarly, avoid treating your hypothesis as “proven” or “disproven” in absolute terms. Many scientific questions are complex, and your study likely contributes one piece to a larger puzzle rather than a final verdict.

5. Citing and Comparing with Published Studies

Do: Situate your findings within the existing literature

A scientific discussion should show how your results align with or diverge from previous studies. Cite key papers that used similar methods or addressed similar questions and compare their findings with yours. You might write:

“These results are consistent with those of Lee et al. (2019), who also found that… In contrast, Smith and Patel (2021) reported no effect of X, possibly because their sample differed in…”

Both confirming and contradictory evidence are valuable. When your results differ from earlier work, take the opportunity to explore possible reasons: differences in sample, context, measures, or analysis.

Do not: Turn the discussion into another literature review

The discussion is not the place to reintroduce all the background literature or to add a long list of new studies. That work belongs in the introduction or a dedicated literature review section. In the discussion, references to other research should serve primarily to support your interpretation and to highlight similarities or differences—not to provide a general overview of the field.

Also avoid letting the ideas of other authors dominate your discussion. Your aim is to show how your study contributes to the conversation, not to restate existing arguments at length.

6. Exploring the Implications of Your Findings

Do: Discuss specific, realistic implications

Readers want to know why your findings matter. In this section of the discussion, you can explore implications for:

• future research (what new questions arise?),
• professional practice (how might clinicians, teachers, engineers, or policymakers use this information?), and
• society more broadly (are there potential benefits or risks?).

Be as concrete as possible. For example:

“If these results are replicated in larger samples, they suggest that incorporating brief mindfulness exercises into routine preoperative care may improve patient outcomes at minimal cost.”

Do not: Exaggerate the impact or speculate wildly

It can be tempting to claim that your findings “revolutionise” practice or “will transform policy worldwide.” Unless such claims are genuinely supported—and they rarely are—it is better to adopt a measured tone. Speculative ideas can be included if clearly labelled as such, but most implications should be closely tied to your data. Realistic, even modest, implications are more convincing than grand promises that readers will doubt.

7. Acknowledging Limitations Honestly

Do: Recognise and explain limitations

No study is perfect, and readers (including reviewers and examiners) know this. Demonstrating that you can critically evaluate your own work is a sign of scientific maturity. When you discuss limitations, consider:

• design constraints (e.g. lack of randomisation, short follow-up, non-representative samples),
• measurement issues (e.g. self-report bias, limited reliability), and
• contextual factors (e.g. single-site data, specific cultural contexts).

Explain briefly how these limitations may affect interpretation, generalisability, or application of your results, and, if appropriate, suggest ways future studies might address them.

Do not: Ignore limitations or let them overshadow your contributions

At one extreme, pretending that your study has no limitations invites scepticism. At the other, focusing so heavily on weaknesses that you undermine your own work can leave readers wondering why the study was done at all. Aim for a balanced approach: be transparent and realistic, but keep the emphasis on what your research does add, even within its constraints.

Also avoid ending the discussion with a long paragraph of limitations. Better to place them near the end but leave the reader with your main conclusions and contributions.

8. Concluding the Discussion Effectively

Do: Provide clear, memorable take-home messages

The final part of your discussion should pull the threads together. In one or two paragraphs, briefly:

• restate your most important findings,
• summarise their main implications, and
• offer a concise overall conclusion.

Many readers (including busy reviewers) turn to the end of a paper to see “what it all means,” so make this section as informative and engaging as possible.

Do not: Inflate or oversimplify

Resist the urge to make your study sound more dramatic than it is. Strong conclusions are not necessarily loud ones; they are clear, honest, and proportionate. Avoid reusing sentences verbatim from earlier sections or simply stitching together phrases from the abstract and introduction. New, well-crafted sentences will leave a better impression and help readers remember your argument.

9. Achieving a Scholarly Style

Do: Write with clarity, balance, and respect

A successful scientific discussion is also a demonstration of your ability to write as a scholar. This means:

• using precise, formal language and complete sentences;
• following a clear line of argument, with logical transitions between ideas;
• citing all sources accurately and treating other researchers’ work with respect.

Confidence is appropriate—you are, after all, the expert on your own study—but avoid arrogance or dismissiveness. Even when earlier studies are flawed, acknowledge their contributions before explaining how your work differs.

Do not: Overload the discussion with jargon or errors

Excessive use of technical terminology and unexplained abbreviations can alienate readers, including those in related disciplines. Use specialist terms when necessary but define them clearly and avoid unnecessary complexity. Finally, a discussion filled with grammatical, spelling, and punctuation errors distracts from your ideas and can make reviewers question your attention to detail.

To avoid these problems, allocate time for careful revision and proofreading. Having a colleague or a professional academic proofreading service check your discussion can help ensure that your argument is not weakened by language or formatting issues.

10. Practical Workflow Tips for Writing the Discussion

Putting all of these elements together can seem daunting. The following practical steps can help:

• Draft a brief outline based on headings such as: summary of main findings, interpretation, comparison with literature, implications, limitations, and conclusions.
• Write freely at first, focusing on content rather than word count.
• Then revise to remove repetition, tighten arguments, and ensure each paragraph has a clear purpose.
• Check that every major point in your introduction is revisited and resolved.
• Ensure that claims are in line with your data and that all key references are cited correctly.

Conclusion

The discussion section is where your research paper comes to life. It is your chance to show readers not just what you found, but what those findings mean, why they matter, and how they fit into the broader scientific story. By following the dos and don’ts outlined here—restating results selectively, linking back to your introduction, interpreting carefully, situating your work within the literature, exploring realistic implications, acknowledging limitations, and crafting thoughtful conclusions—you can write a discussion that is both rigorous and compelling.

Combined with meticulous editing and human proofreading, a well-structured discussion will help ensure that reviewers, editors, and readers understand and appreciate the true value of your research.