Discussion interprets your findings, conclusion answers the research question, and recommendations state proportionate next steps. Your department may place these in Chapter 5, combine them with results or use another structure. If your supervisor has written "this is just a summary" or "conclusion not supported by data" beside your draft, this is the chapter to fix.
The strongest closing chapter stays tied to the results you can show. UniDraft helps organise your confirmed findings and instructions into an editable project document. Start your project write-up, then verify every interpretation and recommendation against your evidence. See the project workflow; it does not carry out the analysis for you.
The short answer
- Discussion: interpret each major result and compare it with the studies in your literature review. Agree, disagree, explain.
- Conclusion: answer each research question or objective from Chapter 1, in order, using only what Chapter 4 showed.
- Recommendations: say who should do what, based on which finding. Keep it to a few.
- Limitations: name the real ones specifically and say how they may have affected the findings.
- Tense: report your own findings in the past tense. "Will show" is wrong for finished work.
Three sections, three different jobs
Hold these apart before you write a word:
- Discuss: interpret your results and relate them to the literature in Chapter 2. Where do you agree with earlier studies, where do you disagree, and why might that be?
- Conclude: answer your original research questions and objectives directly, based only on what your results showed. No new comparisons, no new literature.
- Recommend: give specific, actionable suggestions to specific audiences, grounded in what you found, not generic advice that could close any project.
Most drafts collapse these into one long paragraph that restates the results, gestures at "further studies", and calls it done. Keeping them as separate sections, even under separate headings, is usually the biggest single improvement you can make. Some departments also want a "Summary of Findings" at the start of the chapter; if yours does, keep it to a short recap and do not let it eat the discussion.
Keeping those purposes clear helps the reader trace your reasoning. The actual marking criteria come from your department. The same idea appears in general university writing guidance, which asks that reiterating your findings and interpreting them stay distinct parts of the paper, and that no new results be introduced in the discussion.
Before you write: the setup that makes Chapter 5 easy
Chapter 5 is written last but built from earlier chapters, so gather these first:
- Your list of research questions and objectives, copied word for word from Chapter 1.
- Your results, reduced to one plain sentence per objective ("Objective 2: 41% of respondents reported X").
- The five to ten most relevant studies from Chapter 2, with what each found.
- Anything that went wrong or was not achieved (a low response rate, a failed test, a component that did not work).
If you cannot write a one-sentence result for an objective, the problem is in Chapter 4, not Chapter 5. Go back to how to write the results chapter and fix it there, because a discussion cannot interpret a result that was never clearly stated. Also check that your test choice suited your data, since a conclusion built on the wrong test will not survive a defence panel.
The discussion: relate your results back to Chapter 2
A discussion paragraph should do more than repeat a result. It should say what that result means next to what other researchers found. An illustrative example (the studies and numbers are invented to show the shape):
The finding that 68% of respondents adopted the platform within the first month aligns with Okafor (2023), who similarly found that ease of onboarding was the strongest predictor of early adoption among Nigerian university users. However, it contradicts Bello (2024), whose study reported cost as the dominant barrier, a difference that may reflect the free trial period this study's platform offered, which Bello's did not.
That paragraph states the result, names a prior study it agrees with and why, names one it disagrees with, and offers a plausible reason for the disagreement. Do this for each major finding, not only the headline one. If you need a real source to compare against rather than citing from memory, the free citation finder helps you locate ones you can check.
A repeatable paragraph pattern
Use this order for every finding, and the discussion nearly writes itself:
- Result: one sentence, with the number or theme from Chapter 4.
- Meaning: what it suggests, in plain words.
- Comparison: agree or disagree with named studies from Chapter 2.
- Reason: the most plausible explanation for the agreement or gap.
- Weight: how confident you are, given your sample or method.
Weak versus strong: the same result
Weak (repeats, does not interpret): "Table 4.3 shows that 72% of students used their phones for study. This is a high percentage."
Strong: "Most respondents (72%) used their phones for study, consistent with the pattern in [Author, Year], who linked phone use to limited access to shared computer labs. Since 58% of this sample lived off campus, where lab access is harder, the same explanation is plausible here, although this study did not test it directly."
Notice the strong version admits what was not tested. That honesty is what separates interpretation from decoration.
When your results disagree with expectations
A surprising result is material for discussion, not an embarrassment. If your hypothesis was rejected, or the literature said one thing and your data said another, do three things: state it plainly, offer two or three checkable reasons (sample, place, time, instrument), and say which you find most likely and why. What you must not do is quietly change the objective, drop the inconvenient result, or write a conclusion that pretends the data agreed. A supervisor can compare Chapter 4 with your conclusion in a minute, and a mismatch is one of the fastest things to be flagged.
For engineering, computing and science projects with experiments or a built system, the same logic applies with different evidence: compare your measured performance with the design targets or with published benchmarks you can cite, explain any gap (equipment tolerance, supply variation, dataset size), and say what you would change. If a component did not work as intended, say so and say why. Do not describe as working something you did not demonstrate.
The conclusion: answer your own questions, nothing more
Go back to the research questions or objectives in Chapter 1 and answer each directly, in order, using only what your results support. If you stated three objectives, a reader should find three clear answers. Do not introduce a new comparison, statistic or citation here; if it was not in your results or discussion, it does not belong. Guidance from university libraries makes the same point: don't surprise the reader with information in the conclusion that appeared nowhere else in the paper (see the conclusion guidance).
A simple template, one paragraph per objective:
Objective 1 was to determine the level of ... The study found that ... Objective 2 was to examine ... The results showed that ...
Then close with one short paragraph on the overall contribution: what the study adds, in one or two sentences, without exaggeration.
Watch for these overreaches, all of which draw corrections:
- Causal claims from correlational data. "X causes Y" when you only showed that they move together.
- Generalising beyond your sample. Concluding about "Nigerian students" from 60 students in one department.
- Answering a question you did not ask. A conclusion about cost when no objective or result dealt with cost.
- Certainty language. "This proves" where "this suggests" is what the evidence allows.
Your abstract should agree with this section, since it repeats the key findings and conclusion in miniature; write it after Chapter 5 so the two match.
The tense mistake that quietly undercuts a finished study
A common mistake is easy to fix once you know to look for it: writing a completed study's Chapter 5 as if the work has not happened yet.
Wrong (future tense, as if the study is still ahead of you): "The results will show that mobile banking adoption is influenced by trust in the platform. This study will conclude that trust-building measures are necessary."
Right (past tense, reporting what was found): "The results showed that mobile banking adoption was influenced by trust in the platform. This study concluded that trust-building measures are necessary."
Your project reports work you already did. "Will show" and "will conclude" describe something that has not happened, which weakens a chapter whose point is to report what did. A quick guide to the tenses you will actually use:
| What you are saying | Tense | Example |
|---|---|---|
| What your own study did or found | Past | "Respondents rated trust highest." |
| What a specific earlier study found | Past | "Okafor (2023) reported that..." |
| A supported general statement | Present | "This model describes an association." |
| A limitation you met | Past | "Data were collected from one faculty." |
| What should be done next | Present or modal | "Banks should publish..." / "Future studies could..." |
The "established fact in the present tense, specific studies in the past" split is standard advice in university writing guidance. Conventions vary a little between departments and disciplines, so if your supervisor prefers a different pattern, follow them, and stay consistent. Future tense should not misdescribe completed results; it can still describe genuinely proposed future work. Review the actual output yourself; no automated check guarantees that every tense or claim is correct.
Writing limitations honestly, not defensively
Many students treat the limitations section as a confession to minimise. "Time constraints affected the study" says nothing a reader can use. A specific, honest limitation reads better and costs you nothing:
Vague: "The study was limited by time and resources."
Specific: "Data collection was restricted to students in one faculty because of the time available, which may limit how far these findings generalise to students in other faculties."
A useful test for each limitation: what was it, why did it happen, and how might it have changed my results? If your sentence answers all three, it is a real limitation. If it answers none, delete or rewrite it. Writing on study limitations from university libraries adds that limitations are reported in the past tense (they were discovered after the research), should not be used as an excuse for a thin study, and should not be buried mid-section (see the limitations guidance). Also note that an unexpected or negative result is not a limitation; it is a finding.
Typical genuine limitations for Nigerian undergraduate projects:
| Kind | Honest way to state it |
|---|---|
| Sample | "Only 64 of 100 questionnaires were returned, so results may over-represent respondents who were easier to reach." |
| Scope | "The study covered one department, so it cannot speak for the whole university." |
| Instrument | "Self-reported answers may be affected by respondents' wish to appear favourable." |
| Equipment or data | "Readings were taken with a handheld meter of stated tolerance, so small differences should be treated cautiously." |
| Time | "Observation lasted six weeks, too short to see seasonal changes." |
Every study has limitations of this kind. A study that claims none usually reads as less rigorous, not more.
Recommendations: specific, and for someone
A recommendation should say who should do what, based on which finding. "More research should be done" is a placeholder, not a recommendation. Tie each one to a specific finding and audience (your department, the organisation you studied, a regulator, future researchers) and keep the list short: two or three well-grounded recommendations beat six generic ones.
| Generic (weak) | Grounded (strong) |
|---|---|
| "The government should do more." | "Since 61% of respondents cited unreliable power as the main barrier, the faculty should install backup power in the computer lab." |
| "Students should read more." | "Since only 12% used the library's e-resources, the library should run a short orientation on them in first-year." |
| "Further studies are needed." | "A repeat of this survey across three faculties would test whether the pattern holds beyond Faculty of Science." |
Keep a clear line between recommendations (what others should do with your findings) and suggestions for further study (what researchers could investigate next). Some departments want them as separate headings. Also state your contribution to knowledge in a sentence or two if your department asks for it, and avoid claiming a "novel" contribution unless your literature review actually supports that.
How supervisors flag Chapter 5, and what to do
| Supervisor's comment | What it usually means | Fix |
|---|---|---|
| "This is just a summary." | The discussion repeats results without comparing them | Add named comparisons from Chapter 2 and a reason for each agreement or gap |
| "Conclusion not supported." | A claim in the conclusion has no result behind it | Trace the claim to a table; soften or remove it |
| "Where is objective 3?" | An objective has no answer | Add a direct answer, or revise the objective honestly in consultation |
| "Too generic." | Recommendations could fit any project | Attach each to a finding and an audience |
| "New information." | Something appears for the first time in the conclusion | Move it to the discussion, or cut it |
| "Check your tenses." | Future tense for finished work | Convert to past tense, as in the table above |
When corrections arrive, work through them one by one and keep the earlier version so you can compare; managing final year project supervisor corrections walks through that process. Ask your supervisor whether they want a page limit for the chapter too, because they may have their own preference. If a chapter is failing repeatedly, read what happens if you fail your final year project so you know the realistic options and can act early.
Using Chapter 5 at your defence
Your closing claims are useful material to rehearse for the defence. Be ready to state each conclusion in a sentence, name the result behind it, and explain each limitation without defensiveness. Expect questions such as "what would you do differently?" and "how far can these findings be generalised?", and your limitations and recommendations sections are the answers. Common defence questions and how to answer them covers more.
Pre-submission checklist for Chapter 5
- Each objective from Chapter 1 has its own clearly labelled answer in the conclusion.
- Every conclusion traces to a specific table, figure or result in Chapter 4.
- Relevant findings are compared with appropriate literature or benchmarks where the question and evidence call for that comparison.
- Disagreements with the literature or your hypothesis are explained, not hidden.
- No new results or new citations appear in the conclusion.
- Own findings are in the past tense; no "will show" or "will conclude".
- Limitations are specific, and each says how it may have affected findings.
- Recommendations are few, each tied to a finding and an audience.
- The abstract matches your conclusions.
- Every source you cited appears in your references (how to write the references list).
- Font, spacing and margins meet your university's rules; the free formatting checker tests a .docx against them.
Where UniDraft fits in
Check the current price and included correction allowance in the document workflow before payment.
Sources
- USC Libraries: Writing the discussion section (purpose, verb tense, no new results in the discussion, acknowledging limitations)
- USC Libraries: Writing the conclusion (no new information in the conclusion, recommendations)
- USC Libraries: Writing study limitations (what to include, past tense, placement)