A conclusion tells the reader what your work found and why it matters, in fewer words than any earlier part, and it brings no new results. After that, a lot depends on what kind of writing you are ending. In a research paper the conclusion gives the answer and its importance in a few paragraphs only. A literature review ends by summing up what is already known, naming the gap and moving into your own study. In an Indian PhD thesis the final chapter takes up each objective one by one. Then it gives the contribution, recommendations, limitations and scope for further research.
For each of these three we give one short made-up example. Then we look at the two parts of a thesis ending which examiners read with the most suspicion, recommendations and future scope. Interpreting the results is the work of the discussion chapter, and for that there is its own guide. Last reviewed September 2026.
What is the difference between a paper, literature review and thesis conclusion?
One rule is common to all three, and that is no new evidence. Beyond that they differ in length and in where they point the reader.
| Research paper | Literature review | Final thesis chapter | |
|---|---|---|---|
| Its job | State what the study found and why that matters | Say what is known and what is not, and where your study fits in | Answer each objective, state the contribution, and pass the work on to future researchers |
| Typical length | One to three paragraphs, often 150–300 words | One to two pages in a thesis, and a few sentences in a paper | Often 10–30 pages |
| Points towards | Implications and one or two next steps | Your own objectives | Recommendations and scope for further research |
| Usual heading | “Conclusion(s)”, sometimes a subsection of the discussion | “Summary and research gap” | “Summary, Findings, Suggestions and Conclusion” or “Conclusion and Future Scope” |
These lengths come from what I see in practice. They are not rules. Your journal’s author guidelines and your university’s thesis format take first place. Some departments have their own habit, for example a separate “Suggestions” chapter, and if yours has one, follow it. Before planning your own chapter, take three recent theses of your department from Shodhganga and read last chapter of each.
How do you write the conclusion of a research paper?
Keep this one short. The generic guide for authors from Elsevier says the key conclusions of the study “may be presented in a short Conclusions section, which may stand alone or form a subsection of a Discussion or Results and Discussion section.” Check your target journal first and then decide.
How is it different from the discussion and the abstract?
In the discussion you argue and discuss about other studies. Your numbers sit next to theirs there, and you explain anything unexpected. When you reach the conclusion that arguing is finished, and you only report where it ended.
The abstract summarises the full paper, method also, for a person who is still deciding whether to read it. The conclusion is written for a person who has already read it. So there is no need to give the sample size again there. What it needs is the answer, why that answer matters, and where the work should go next. Try one check. If somebody swapped your abstract and conclusion and nobody would notice, rewrite the conclusion (the abstract guide covers the other side).
ICMJE wrote its recommendations for medical journals. Still I would give the same advice to a commerce scholar. Authors are asked to connect the conclusions with the goals of the study. And they should not make unqualified statements, or draw any conclusion which their data do not support.
What should you leave out?
- Any result that is missing from the results section. Put it in the results chapter, then mention.
- The abstract copied once more with only the tense changed.
- New citations. The literature was already handled in the discussion.
- Claims bigger than the design can support. A survey done in one district cannot tell what “Indian farmers” do, and a correlation cannot show a cause.
- Big closing lines about national development or the need for “more awareness”.
Example
The study and all its numbers are invented. Take a civil engineering paper that tested bricks made with different proportions of fly ash.
Bricks in which 30 per cent of the clay was replaced by fly ash met the strength requirement we tested against, whereas bricks with 40 per cent or more did not. For small kiln owners located near thermal power plants, this points to a practical upper limit on fly-ash use without any change to the firing process. We took ash from one source only and fired all bricks at a single temperature. Ash from other plants needs trying next, and durability under repeated wetting and drying also needs testing.
See what that paragraph carries. The answer comes first, then why kiln owners should care, then where the evidence stops. No citation or p-value is there.
How do you write the conclusion of a literature review?
Here there are three moves, and the order matters. First pull the themes together into one statement of what is known. Next say what is still not known. Last, go directly into your own study. Scholars mostly forget this last move. The chapter should finish on something very near to your objectives, so the methodology chapter can open by explaining how you studied them.
A full sample closing paragraph for a management thesis is already there on our literature review examples page, so we are not repeating it here. This one is shorter, from agriculture. The topic is invented, and citations are removed deliberately.
The studies reviewed agree that drip irrigation reduces water use on sugarcane farms, and most of them report higher yields. However, nearly all of them measure outcomes only in the first two seasons after installation, on farms that received a subsidy and technical visits. Almost nothing is known about whether smallholders continue to use drip systems after that support ends. The present study therefore follows adopters in two districts of Maharashtra over three seasons to identify the factors associated with continued use.
Read it again and see what is missing. No new study comes in. Each claim summarises work that was discussed earlier in the chapter, and there is no list of author names. The gap is a specific one (continued use after the support ends). Compare that with “no study has been done in this area”, a line examiners have learnt not to trust. In a journal article the same three moves become two or three sentences just before your aim.
What about the conclusion of a research proposal?
A synopsis closes in the same way, only it faces forward. It states the problem in one sentence, then the objectives and the contribution you expect. Do not promise findings. Many scholars are writing “The study will show that…” and then a committee member asks them, how do you know before doing it? The other parts of a synopsis we have covered in the research proposal guide.
How do you write the final chapter of a PhD thesis?
Many Indian theses in commerce, management, education and the social sciences give the last chapter a title like “Summary, Findings, Suggestions and Conclusion”. In engineering and science, “Conclusion and Future Scope” is more common. What goes inside is quite similar. The thesis structure guide explains where the chapter comes in the thesis, and here we look only at its contents.
Brief recap of the study
One or two pages, in past tense, on the problem, the objectives and the method.
Findings, objective by objective
A separate sub-section for each objective, worded exactly as in Chapter 1, with a pointer to the table or section where that result is reported.
Conclusions and contribution
What the findings add up to, and what the thesis contributes to theory, method or practice.
Suggestions or recommendations
Each one addressed to a named actor and traceable to a finding.
Limitations
Keep it brief, or point back to the discussion if they are already covered there.
Scope for further research
Three to five specific studies that someone could design from what you have written.
When a final chapter comes back for rework, mostly it is the findings section, and mostly for one and the same problem. The scholar copies the results chapters once again, table after table. Examiners have already read those tables. Now they want each objective in the Chapter 1 wording, with a direct answer under it. If your objectives have drifted since the synopsis, correct that first (the objectives guide shows how to word them so that they can be answered).
Findings by objective: an illustrative extract
Suppose there is a commerce thesis on UPI use by small retailers in two districts of Tamil Nadu. All of it is invented.
Objective 2: To identify the factors associated with regular use of UPI by small retailers.
Regular use was most strongly associated with in-person onboarding by a bank or payments agent, and next with the share of customers under 35 (Section 5.3, Table 5.7). Once these were controlled for, shop size and the owner’s education were not significant. Among cash-only retailers, the reason given most often was fear of disputes over failed transactions (Section 5.5).
Each objective gets one short block and one pointer back to the evidence. Do this for every objective and the examiner can tick them off in five minutes.
After that, take a step back. In this thesis the conclusion could be a single paragraph. It would argue that adoption depended more on the way retailers were introduced to UPI than on who those retailers were. The contribution then says what this adds. Maybe it corrects earlier researches that stress education and income, or maybe it gives a tested questionnaire which others can use again. Claim only what you are able to defend. At the viva, “What is your contribution?” is a common opening question, and you will find the usual follow-ups in the viva questions guide.
How should recommendations follow from the findings?
Each recommendation has to point back to some finding. Which finding gave you this idea? No answer, then cut it, even if it sounds sensible.
Many times this is the place where examiners write “generic” in the margin. You must have seen lines like “the government should create awareness” or “training programmes should be conducted” many times. Both can be written without collecting even one questionnaire, so they are weak.
| Finding (illustrative) | Weak recommendation | Stronger recommendation |
|---|---|---|
| Retailers who were onboarded in person by a bank or payments agent were much more likely to be regular users. | The government should create more awareness about digital payments. | Banks and payment providers in the two districts could hold in-shop demonstrations for kirana owners, because in-person onboarding was the factor most closely linked with regular use. |
| The reason heard most from cash-only shops was fear of disputes when a transaction fails. | Retailers should be educated about UPI. | Payment providers could show the complaint and refund route in Tamil, inside the merchant app itself. |
The stronger versions name who should act and what they should do, and the reason travels along with them. They also don’t go beyond the sample, which is two districts and not “all of India”.
Six well-grounded suggestions are better than twenty. At the pre-submission seminar, the committee will pick your weakest one and question you on that only.
How do you write the scope for further research without being generic?
“Future research may be carried out in other states with a larger sample.” I have seen this sentence in more theses than I can count. It tells the next scholar nothing useful.
Good future scope comes out of your own limitations, your unexpected results and the new questions your findings opened up. Every suggestion should say what to study, and with whom or with what. It should also say why your work makes that study worth doing.
Generic: The study can be extended to other districts and other types of digital payment.
Specific: In-person onboarding was the strongest correlate of regular use, but the design was cross-sectional. To know if this link is causal, somebody can do a field experiment where only a randomly chosen half of the retailers get onboarding visits. Education made no difference once onboarding was controlled for, and that finding also deserves a qualitative study with low-literacy retailers on how they learn to use the app.
Engineering is no different. A line like “The model can be improved using deep learning” is generic. Compare it with “The model was tested only on summer data from one plant, so monsoon data is the next test”, which gives somebody a project.
Want a specialist to help with this in your own work? Thesis Chapter Review and Feedback: a free first consultation, then a written scope and fixed quote.
Sources
FAQ
Questions scholars ask
Can I cite references in the conclusion?
In a paper, better to avoid it. A thesis final chapter can carry the odd citation when you are stating the contribution. Only be careful that the chapter does not become a second literature review by the end.
Should the conclusion be written in past or present tense?
Use past tense for what you did and what you found. For what it means, present tense works (“this suggests…”). This is the same convention you follow in the discussion.
Are “suggestions” and “recommendations” the same thing?
Indian theses use the two words almost interchangeably. Check the recent theses of your department and go with whichever heading they use.
How long should the final chapter of a thesis be?
No fixed rule is there. Mostly it is shorter than the results chapters. If yours has become longer than the methodology chapter, probably it is repeating the results.
