← how-to

How to Write a Discussion Section: Dissertation, Paper and Lab Report Guide

To write a discussion section, answer your research question, interpret each finding, compare it with the literature, then set out limitations, implications and future research. Here is a 7-step method with examples.

16 November 2026
discussion sectiondissertationresearch methodslab reportacademic writing

To write a discussion section, start by answering your research question in one or two sentences, then interpret each key finding, explain why it may have occurred and compare it with the studies in your literature review. Finish by setting out the limitations, the theoretical and practical implications, and specific directions for future research, without introducing new results.

This guide works for dissertation and thesis chapters, journal-style papers and lab reports. It covers how the discussion differs from the results and the conclusion, how long it should be, a seven-step method and how to compare your findings with the literature. It also shows how to handle unexpected or non-significant results and how to write limitations without undermining your work. Two annotated examples follow, one quantitative and one qualitative, built from invented studies. The advice comes from FLN Tutors (Freelancer Network) tutors who supervise and mark dissertations and research projects across the sciences, social sciences, health and business.

Key facts: the discussion section at a glance

QuestionShort answer
What is it for?Explaining what your results mean and why they matter
Where does it go?After the results, before the conclusion
How long is it?Often 15 to 25% of a dissertation; one of the longest sections of a paper
What tense?Past for your findings, present for what they mean and for established knowledge
What does it include?Answer, interpretation, comparison with literature, limitations, implications, future research
What does it exclude?New results, raw statistics and new background theory
How is it ordered?Usually the same order as your results or research questions

What is a discussion section, and how is it different from results and conclusion?

The discussion section is the part of a research report where you interpret your findings: you explain what they mean, how they relate to existing research and theory, what limits their reliability and why they matter. The results section reports what you found; the discussion explains it; the conclusion summarises the answer and its significance briefly.

SectionMain questionTypical contentTypical tense
ResultsWhat did I find?Data, statistics, themes, tables and figuresPast
DiscussionWhat does it mean, and why?Interpretation, comparison, explanation, limitations, implicationsPast and present
ConclusionSo what is the answer?Short summary, contribution, recommendationsPresent

Some papers and qualitative dissertations combine results and discussion in one chapter. Even then, keep reporting and interpreting distinguishable, for example by finishing each theme with a short interpretive paragraph. For guidance on the conclusion chapter itself, see our conclusion guide.

How long should a discussion section be?

There is no fixed rule, but these planning ranges are common for student work. Always check your handbook.

DocumentTypical total lengthDiscussion shareApproximate discussion length
Undergraduate dissertation10,000 words15 to 20%1,500 to 2,000 words
Master's dissertation15,000 words15 to 20%2,250 to 3,000 words
Empirical journal-style paper6,000 words20 to 25%1,200 to 1,500 words
Undergraduate lab report2,000 words25 to 30%500 to 600 words

A discussion much shorter than the results usually means you are describing rather than interpreting. One that is far longer than the results may be speculating beyond the data.

How to write a discussion section in 7 steps

  1. Answer the research question. Open with one or two sentences that state the main finding in plain language, without statistics.
  2. Interpret each key finding. Take findings in the order of your results or research questions, and explain what each one shows.
  3. Explain why. Offer plausible explanations, drawing on theory, your method and the context of your study.
  4. Compare with the literature. Say whether each finding agrees with, contradicts or extends the studies you reviewed, and suggest reasons for any differences.
  5. Address unexpected results. Discuss findings that did not match your hypotheses, rather than leaving them out.
  6. State the limitations. Name specific weaknesses, explain how they affect interpretation, and note what you did to reduce them.
  7. Set out implications and future research. Explain what the findings mean for theory, practice or policy, and propose concrete next studies.

Steps 2 to 4 repeat for each finding. The opening answer and the closing implications frame the section.

How do you interpret your findings?

Interpretation means moving from "what" to "so what". For each finding, answer three questions: what does this show about the research question, why might it have happened, and how confident can we be?

Match your verbs to the strength of your evidence. This is called hedging, and it is one of the main markers of academic maturity in a discussion.

Strength of evidenceSuitable verbs and phrasesAvoid
Strong and consistentshows, demonstrates, provides clear evidence thatproves
Moderateindicates, supports the view that, is consistent withconfirms beyond doubt
Tentativesuggests, may reflect, one possible explanation isdefinitely, obviously
Speculativeit is possible that, further research could test whetherstating speculation as fact

Two rules prevent most problems. First, do not restate the results: if a sentence could appear unchanged in the results section, it is not interpretation. Second, do not overreach: a correlation does not show causation, and a sample of first-year students does not represent all students.

How do you compare your results with the literature?

Comparison with previous research is what turns a description into a contribution. Start each comparison with your own finding, then bring in the literature; if you start with other studies, the section begins to read like a second literature review.

Use one of four patterns:

  • Agreement: "This finding is consistent with earlier work showing that..."; then explain what your study adds, such as a new population, setting or method.
  • Contradiction: "Unlike earlier studies, which found..., this study found..."; then propose reasons, such as differences in sample, measurement, context or timing.
  • Extension: "This study extends previous work by showing that the effect also holds for..."
  • New finding: "No previous study reviewed here examined..."; then explain its possible significance cautiously.

Return to the theories and studies from your literature review. Introducing many new sources at this stage suggests the review missed something; a few new citations are acceptable when an unexpected result needs explaining. The University of Guelph's guide to the discussion section shows the same pattern of comparing your own result first.

How do you handle unexpected or non-significant results?

Unexpected and non-significant results are findings, not failures, and markers expect you to discuss them.

  • Report them honestly. Never omit a result because it did not support your hypothesis.
  • Consider methodological explanations: sample size and statistical power, measurement tools, timing, or how the intervention was delivered.
  • Consider substantive explanations: the theory may not apply to this population or context, or a different mechanism may be at work.
  • Distinguish "no evidence of an effect" from "evidence of no effect". A non-significant result in a small sample means the study could not detect an effect of that size, not that none exists.
  • Separate statistical from practical significance. A statistically significant but tiny effect may matter little in practice; a non-significant but sizeable effect in a small sample may deserve further study.

How do you write the limitations?

Limitations are the specific features of your design or data that restrict how far your conclusions can be trusted or generalised. Written well, they strengthen your credibility, because they show you understand your own study. Our methodology guide explains how to justify design choices earlier in the dissertation, so the limitations section should not come as a surprise.

For each limitation, write three things:

  1. The limitation: for example, a convenience sample from one university.
  2. Its effect: results may not generalise to students at other institutions or in other countries.
  3. The mitigation or response: what you did to reduce it, or how future research could address it.
Weak limitationStronger version
"The sample size was small.""With 24 participants, the study could detect only large differences between groups, so the absence of an effect on attendance should be treated as inconclusive."
"There may have been bias.""Because participants self-reported their study hours, the totals may be inflated by social desirability; future studies could use learning platform logs."
"More research is needed.""A follow-up with students at a non-selective university would test whether the effect holds where entry grades are more varied."

Choose the two to four limitations that matter most for your conclusions. A long list of generic weaknesses suggests that you do not know which ones are important.

How do you write implications and future research?

Implications explain why your findings matter. Divide them into two kinds:

  • Theoretical implications: whether your results support, challenge or refine an existing theory or model.
  • Practical implications: what practitioners, organisations or policymakers could do differently, stated with appropriate caution.

Recommendations for future research should follow directly from your limitations and from new questions your findings raise. Make them concrete: name the population, design or variable a future study should use. "Future research should explore this further" earns no credit; "a longitudinal study following the same cohort into the second year would show whether the effect persists" does.

Annotated example: a quantitative discussion paragraph

The study and the citations below are invented to show the technique; they do not refer to real publications.

Scenario: 212 first-year students were split equally between a weekly online quiz group and a rereading group (106 each). Quiz users scored a mean of 64.2 on the module exam and rereaders 58.7, a difference of 5.5 points. With a pooled standard deviation of 13.4, the results section reported t(210) = 2.99, p = .003, d = 0.41.

"Students who completed weekly online quizzes scored higher in the final exam than those who reread their notes. [Answer first, no statistics] The difference of 5.5 points corresponds to an effect size of 0.41 standard deviations, which lies between conventional benchmarks for small (0.2) and medium (0.5) effects and is large enough to matter for students close to a grade boundary. [Interpretation and practical significance] This is consistent with laboratory research on retrieval practice (Hale, 2019) and extends it to a full-semester module with real assessment stakes. [Comparison and extension] However, the effect was smaller than the 0.6 standard deviations reported in a short classroom study (Osei and Varga, 2022), possibly because quiz completion was voluntary and some students in the quiz group completed few quizzes. [Contradiction and explanation] Because students chose their own group, more motivated students may have selected quizzing, so the result shows an association rather than a causal effect. [Limitation] A randomised design with required quiz completion would test whether the effect is causal. [Future research]"

Notice that the paragraph uses the statistics only to interpret magnitude; the full statistical reporting stays in the results section.

Annotated example: a qualitative discussion paragraph

This example also uses an invented study and invented citations.

Scenario: semi-structured interviews with 14 hospital nurses about night-shift work produced a theme labelled "invisible recovery", describing how nurses used rest days to recover rather than for personal life.

"The theme of invisible recovery suggests that night-shift fatigue extends well beyond the shift itself, consuming time that rostering systems count as rest. [Interpretation] This adds a lived-experience dimension to survey research reporting high fatigue among night-shift staff (Lindqvist, 2018), by showing how nurses themselves frame rest days as recovery time rather than free time. [Comparison and extension] Unlike earlier accounts that located the problem mainly in individual sleep habits (Daley, 2020), participants here linked recovery to roster patterns, particularly quick returns between shifts. [Contrast] The sample came from a single trust, and nurses who volunteered may have felt more strongly about fatigue than their colleagues, so these findings indicate how fatigue can be experienced rather than how common it is. [Limitation framed for qualitative research] For practice, the findings support reviewing quick-return shift patterns; future research could compare trusts with different rostering policies. [Implications and future research]"

Qualitative discussions avoid claims about frequency or generalisation and talk instead about transferability: whether the insights might apply in similar settings.

How does a discussion differ in a dissertation, a journal paper and a lab report?

FeatureDissertation discussionJournal-style paperLab report
LengthA full chapter, often with subheadingsA section of several paragraphsA few paragraphs
Literature comparisonExtensive, tied back to the literature review chapterFocused on the most relevant studiesBrief, often against expected or textbook values
OrganisationBy research question, hypothesis or themeBy key findingBy aim, then sources of error
LimitationsA dedicated subsectionA paragraph near the endError analysis and uncertainty
ImplicationsTheory, practice and future research in depthBrief, matched to the journal's audienceImprovements to the experiment
ConclusionUsually a separate chapterShort closing paragraph or separate sectionA short conclusion statement

In lab reports, discussion centres on whether results match theory or accepted values, the size of experimental error and how the method could be improved. Our lab report guide covers error analysis and uncertainty in detail.

How do discussion conventions differ in the UK, US and elsewhere?

  • UK and Ireland: undergraduate and master's dissertations usually have a separate discussion chapter followed by a conclusion chapter, and markers reward critical engagement with the literature and well-reasoned limitations. Loughborough University London's guide on writing a discussion reflects typical UK expectations.
  • United States: master's theses and doctoral dissertations often follow a five-chapter model in which chapter 5 combines discussion, conclusions and recommendations. Psychology and education papers follow APA guidance, which expects limitations and generalisability to be addressed in the discussion.
  • Australia and New Zealand: honours and master's theses follow a UK-like model, and many universities accept thesis-by-publication formats with a general discussion chapter linking the papers.
  • Europe (ECTS): master's theses in the Netherlands, Germany and Scandinavia often follow journal article structure, with a discussion section rather than a chapter, and article-based PhDs end with a general discussion.
  • Singapore and Malaysia: universities follow British or international journal conventions, often with a combined discussion and conclusion chapter in master's projects.
  • Gulf states: institutions follow the model of their partner system: US-style five-chapter theses at American-model universities and UK-style chapters at British branch campuses.

What do markers look for, and what mistakes cost marks?

Markers look for five things in a discussion: a clear answer to the research question, interpretation rather than repetition, critical comparison with the literature, honest and specific limitations, and implications that follow from the evidence.

Common mistakes:

  1. Repeating the results. Statistics and descriptions without interpretation.
  2. Ignoring the literature review. Findings discussed in isolation.
  3. Overclaiming. Causal language from correlational data, or generalising beyond the sample.
  4. Leaving out awkward results. Markers notice when a reported finding disappears.
  5. Generic limitations. "The sample was small" without saying what that means.
  6. New results. Data or analyses that never appeared in the results section.
  7. Vague future research. "More research is needed" with no design or question.
  8. Ending on limitations. Finish on implications or contribution so the section ends on what the study adds.

To see how a strong dissertation mark affects your overall result, try our degree classification calculator.

Getting help with your discussion section

If your supervisor says your discussion is "too descriptive" or "not critical enough", a tutor can help you see the difference between reporting and interpreting. Through our dissertation support, you can get feedback on a draft discussion chapter, help linking findings back to your literature review, and guidance on writing limitations and implications. We match you with a tutor who holds a postgraduate degree in or near your field and has supervised or marked dissertations that use your type of data.

Tutors explain and give feedback; the interpretation and writing in your dissertation must be your own.

Frequently asked questions

What tense should the discussion section be written in?

Use the past tense for what your study found, for example "participants reported", and the present tense for what the findings mean and for established knowledge, such as "this suggests" or "previous research shows". Use the present or future for implications and recommendations. Keep tense consistent within each paragraph so the reader can follow the shift.

How do I start a discussion section?

Start with a direct answer to your research question or main hypothesis in one or two plain sentences, without statistics. For example, "Weekly quizzes were associated with higher exam scores than rereading." Then move into interpretation of each finding. Avoid opening with a long recap of the aims or methods, which the reader already knows.

Can I cite new sources in the discussion?

Yes, a few. Most comparisons should draw on the studies in your literature review, but it is acceptable to cite new sources when an unexpected finding needs explaining or a new angle emerges. If you need many new sources, consider whether your literature review should be updated to include that body of work.

Should the discussion follow the same order as the results?

Usually yes. Following the same order, whether by research question, hypothesis or theme, lets the reader connect each finding to its interpretation. You can group closely related findings, but avoid jumping around. Start with the findings most central to your research question and leave secondary or exploratory results until later.

Can I include new results in the discussion?

No. Every finding you interpret must already appear in the results section. If you notice something important while writing the discussion, add it to the results first and then discuss it. The only numbers that belong in the discussion are those you need to describe a finding's magnitude or compare it with other studies.

How do I discuss non-significant results?

Report them as findings, then consider why the study did not detect an effect: sample size and statistical power, measurement, timing or the theory itself. Explain that a non-significant result is not proof that no effect exists. Where available, discuss the effect size and confidence interval, which show what size of effect the data can rule out.

How many limitations should I include?

Most student dissertations discuss two to four limitations in depth, chosen for their effect on the conclusions. Each should state the limitation, its likely effect on the findings and how it was reduced or could be addressed. A long list of minor issues weakens the section, because it suggests you cannot tell which limitations really matter.

What is the difference between implications and recommendations?

Implications explain what your findings mean for theory, practice or policy: why they matter. Recommendations are specific actions that follow from those implications, such as changing a procedure or conducting a particular study. In many dissertations, implications sit in the discussion and recommendations appear in the conclusion chapter, but departments vary.

How much of the discussion should compare with other studies?

There is no fixed share, but in most dissertations comparison with the literature runs through every finding rather than sitting in a separate block. Each key finding should connect to at least one relevant study or theory. If whole paragraphs mention no other research, your discussion may be too descriptive; if they mention nothing else, it may lack your own interpretation.

Is the discussion the longest chapter of a dissertation?

It is often among the longest, alongside the literature review, and typically takes about 15 to 20% of an empirical dissertation's word count. In qualitative or combined findings-and-discussion chapters, it can be longer. Length should follow the number and complexity of your findings rather than a fixed target.

How do markers grade a discussion chapter?

Markers check whether the discussion answers the research question, interprets rather than repeats findings, engages critically with the literature, identifies specific limitations and draws justified implications. Higher bands reward nuanced explanations, honest treatment of unexpected results and a clear sense of the study's contribution to knowledge or practice.

Can a tutor give feedback on my discussion chapter, and how does it work?

Yes. Share your draft chapter and research question for a free quote, then choose written feedback with delivery options from 3 to 48 hours or book live sessions at the hourly rate shown on the tutor's profile. Free revisions within the original brief are included, everything is confidential, and the dissertation stays your own work.

Want a subject expert to help with this?

Get help with your lab report →

Money-Back Guarantee

Full refund if we miss the agreed deadline, can't find a suitable expert, or major errors can't be fixed by revision.

Learn more →

100% Confidential

Your identity, order details and institution are never shared with any third party.

Learn more →

Deadline-Backed Delivery

Every order has an agreed deadline. If we don't deliver by it, you can claim a full refund.

Learn more →

Get started today

Ready to get expert help?

Tell us what you need and get matched with a tutor who teaches your subject. Delivered on time, every time.

See how it works
Instant quoteNo account needed to get a priceFree revisions100% confidential