Why do examiners so often write “the conclusion merely repeats the discussion” or “the research questions are never answered directly”? Because the thesis conclusion chapter is written last, in the final week, when the author is tired and the discussion chapter has already said everything that felt important. The result is a summary in a new font.
This article is part of our guide How to Write a Master’s Thesis by the Mastermind PhD team and relates to our Master’s Thesis support service. Sources last checked: November 2026.
A conclusion has a different job: it closes the argument opened in chapter one by answering each research question in plain words, stating what the thesis adds, admitting what it could not do, and saying who should do what next. This guide gives the chapter a structure with word shares, shows the moves paragraph by paragraph, and ends with a checklist. The example paragraphs are invented for illustration and marked as such.
What is the conclusion and recommendations chapter? The thesis conclusion chapter is the final chapter, which restates the aim, answers each research question or hypothesis directly, states the contribution to knowledge and practice, acknowledges the main limitations and their effect on the conclusions, and sets out recommendations for practice or policy and for further research. It introduces no new data, literature or arguments; everything in it has already been established in the results and discussion chapters.
What is the thesis conclusion chapter for, and how does it differ from the discussion?
The discussion interprets: it explains what the findings mean, compares them with earlier studies, weighs explanations and qualifies claims. The thesis conclusion chapter decides: it says, in one place and in the words of the original questions, what was found and what follows.
The University of Southampton’s dissertation guide puts it simply: the conclusion “serves to reinforce your main argument and findings, before considering the wider implications of your research”, and it should be treated as a chapter in its own right rather than a paragraph tacked on. If your handbook combines discussion and conclusion into one chapter, the same moves still appear as the final section; our article on the discussion chapter covers the interpretive part.
How long should it be and where does it sit?
Short. Southampton’s rule of thumb is roughly 5 to 10 per cent of the dissertation’s word count, usually nearer 5; for a 15,000-word thesis that is 750 to 1,500 words. The USC Libraries writing guide, written for research papers, describes the conclusion as often the shortest section, requiring only two or three well-developed paragraphs; a thesis needs more, because it has several questions to answer and recommendations to make, but the principle holds. In a six-chapter dissertation the thesis conclusion chapter is chapter six; in the five-chapter model it is the final section of chapter five.
📱 Talk to the Mastermind PhD team — You are told the specialist’s field and degree before work starts, and a meeting is scheduled to review your requirements.
What structure works for a thesis conclusion chapter?
The order below maps onto what examiners look for. The word counts are for a 1,000-word conclusion in a 15,000-word thesis; scale them to your budget.
| Section | What it does | Words |
|---|---|---|
| Restatement of aim and approach | One paragraph: the problem, the aim, the design, in plain prose with no statistics | 100 |
| Answers to the research questions | One short paragraph per question, in the words the question was asked | 300 |
| Contribution | What is now known or available that was not before: theoretical, empirical, methodological or practical | 150 |
| Limitations and their effect | The three or four limitations that bear on the conclusions, each with its effect | 150 |
| Recommendations for practice or policy | Who should do what, within the limits of the evidence | 150 |
| Recommendations for further research | Specific studies that would resolve what this one could not | 100 |
| Closing statement | Two or three sentences that return to the opening problem | 50 |
Headings for each section of the thesis conclusion chapter are common in UK and European theses and help the examiner find the answers; in some disciplines the chapter runs as continuous prose. Follow recent theses from your department.
How do you answer each research question?
Write the question as a subheading or in italics, then answer it in two to four sentences that state the finding, its direction and its strength, without the statistics that already appear in chapter four. The Academic Phrasebank at the University of Manchester groups useful openings under “restating the aims of the study” and “summarising main research findings”: “This study set out to…” and “The results of this investigation show that…” are the tone to aim for.
Illustrative example (fictional study): Research question 1 asked whether structured remote onboarding is associated with lower turnover intention among junior developers in small software firms. It is. Developers who reported a documented onboarding sequence with a named mentor expressed lower intention to leave in their first year than those who did not, and the association held after controlling for tenure and company size. The size of the association was moderate, which means onboarding structure is one factor among several, not a decisive one.
Notice what the paragraph does not do: it does not cite, it does not report r or p, and it does not hedge so heavily that the answer disappears. The hedging was done in the discussion; here, in the thesis conclusion chapter, the author commits to the answer the evidence supports.
How do you state the contribution without overclaiming?
Name the type of contribution and the audience who gains from it. Empirical: a known relationship tested in a new population or setting. Theoretical: a framework extended, qualified or challenged. Methodological: an instrument adapted and validated, a procedure made replicable. Practical: a tool, protocol or recommendation usable by practitioners. Match the verb to the design: a cross-sectional survey “extends evidence” and “suggests”; it does not “prove” or “establish”.
Illustrative example (fictional): The thesis makes two contributions. Empirically, it extends evidence on onboarding and retention, drawn mostly from large on-site organisations, to small remote-first firms, where the association was present but weaker where supervisor support was high. Methodologically, it adapts an existing onboarding-structure scale to remote work and reports its reliability in this population, so that later studies can use a shared measure.
How do you write limitations in a conclusion?
Briefly, and always with the consequence. The discussion chapter already explained each limitation; the thesis conclusion chapter names the ones that constrain the answers and says how far they constrain them. Three parts per limitation: what it is, how it could have affected the findings, how the reader should weigh the conclusions as a result. The USC guide warns against both extremes: negative aspects of the research should never be ignored, but neither should the author apologise for the work or cast doubt on its validity.
Illustrative example (fictional): Two limitations shape the conclusions. The sample was recruited through professional networks, so disengaged developers who had left the industry are under-represented; the association may therefore be weaker in the full population than reported here. The design was cross-sectional, so the direction of the association cannot be settled; the recommendations below are framed as practices to consider, not as proven causes of retention.
How do you write recommendations for practice and for research?
Recommendations for practice in a thesis conclusion chapter name an audience and an action, and they stay inside the evidence. “Organisations should pay more attention to onboarding” is not a recommendation; “Scale-ups without a human resources function should document a four-week onboarding sequence with a named mentor and review it at the end of the probation period” is. Where the design was correlational, use “consider”, “pilot” and “evaluate” rather than “must”. The Phrasebank’s section on “setting out recommendations for practice or policy” offers neutral frames such as “The findings of this study have a number of practical implications.”
Recommendations for research follow from the limitations and the open questions, and they are specific enough for someone to act on: a design, a population, a measure, a time frame. “Future studies should…” followed by “use a longitudinal design measuring actual turnover at twelve months in a random sample of firms registered with a national industry body” is actionable; “more research is needed” is not.
Illustrative example (fictional): For practice, founders of small remote-first firms should consider documenting an onboarding sequence of at least four weeks with a named mentor, and should evaluate its effect on first-year retention before scaling it. For research, a longitudinal study that measures actual turnover at twelve months, rather than intention, in a random sample of firms, would establish whether the association reported here reflects a causal effect of onboarding structure.
Which phrases help, and which hurt?
Helpful openings signal the move you are making: “This study set out to”, “Taken together, these results suggest that”, “Notwithstanding these limitations, the study suggests that”, “Future studies should”.
Phrases to avoid are those the UNC Writing Center and the USC guide both flag: stock openers such as “In conclusion” or “In summary”, which Southampton calls a clichéd approach; stating the thesis for the first time in the conclusion; introducing a new idea or new evidence; restating the thesis with no substantive change; and emotional appeals that clash with the analytical tone of the rest of the thesis. If a sentence in the thesis conclusion chapter would need a citation, it belongs in the discussion or the literature review.
A checklist before you submit the thesis conclusion chapter
- Each research question from chapter one is answered in its own words, in the order asked.
- Every hypothesis stated in chapter one is reported as supported, partly supported or not supported.
- The contribution is named by type and audience, with verbs that match the design.
- Limitations appear with their effect on the conclusions, not as a list of confessions.
- Each recommendation names who should do what, and stays within the evidence.
- No new data, sources, tables or arguments appear.
- No statistics are repeated; the reader is pointed to chapter four for them.
- The chapter is within 5 to 10 per cent of the word budget and ends by returning to the opening problem.
- The abstract, introduction and conclusion agree on the questions, the findings and the contribution.
Read chapter one and the thesis conclusion chapter back to back before submission. If a question was asked and not answered, or a significance was promised and not delivered, fix it in the conclusion rather than in the introduction. A reviewer can check the chapter for alignment, repetition and structure; the answers, the claimed contribution and the recommendations are the author’s own, and they are what the examiners will ask you to defend.
📱 Ask on WhatsApp now — 5 years of experience across all disciplines and 14 academic services. Two-stage payment (80/20); revisions are free until supervisor approval, within 30% of the chapter.
Frequently asked questions
How long should a thesis conclusion be?
About 5 to 10 per cent of the thesis, usually nearer 5: roughly 750 to 1,500 words for a 15,000-word master’s thesis, and proportionally more for a doctoral thesis with several research questions. If the thesis conclusion chapter is longer than the results chapter, it is probably repeating the discussion.
Should the conclusion repeat the key statistics?
No. State the findings in words, with their direction and approximate strength, and leave the statistics in the results chapter. A reader who wants the numbers knows where to find them; an examiner who sees them twice reads the chapter as padding.
Can I include new recommendations that did not come from my findings?
No. Every recommendation must trace back to a finding or a limitation reported earlier. A recommendation that the data do not support, however sensible, reads as opinion and invites the question “where in your results is this?”
Is it acceptable to write the conclusion in the first person?
If your department allows the first person elsewhere in the thesis, yes; “I conclude that” is clearer than “it is concluded that”. Use one convention consistently across all chapters and check the handbook, because some disciplines still prefer impersonal constructions.
What is the difference between implications and recommendations?
Implications say what the findings mean for theory, practice or policy; recommendations say what someone should do about them. Implications usually sit in the discussion, recommendations in the conclusion, and each recommendation should be traceable to an implication and to a finding.
Sources
- University of Manchester. Academic Phrasebank: Writing Conclusions. https://www.phrasebank.manchester.ac.uk/writing-conclusions/ (accessed October 2026)
- University of Southampton Library, Academic Skills Service. Writing the Dissertation: Conclusion. https://library.soton.ac.uk/writing_conclusiondissertation (accessed October 2026)
- USC Libraries. Organizing Your Social Sciences Research Paper: 9. The Conclusion. https://libguides.usc.edu/writingguide/conclusion (accessed October 2026)
- The Writing Center, University of North Carolina at Chapel Hill. Conclusions. https://writingcenter.unc.edu/tips-and-tools/conclusions/ (accessed October 2026)
- Murray, R. (2017). How to Write a Thesis (4th ed.). Open University Press.
- Evans, D., Gruba, P., & Zobel, J. (2014). How to Write a Better Thesis (3rd ed.). Springer.
Sources last checked: November 2026. This article is general educational guidance and does not replace your university regulations or your supervisor’s instructions. How we prepare our content.