Short answer: the discussion interprets — it weighs what the findings mean, sets them against prior work, admits limits and hedges every verb to the design. The conclusion resolves — it answers the research question in one breath, folds the findings into a single pattern and stops.
Both sections sit at the end of the paper, describe the same findings and lean on the past tense — which is why the boundary blurs exactly where marks are cheapest to lose. A conclusion that re-opens the argument reads as padding; a discussion that never resolves reads as unfinished. It is a job-description problem, and job descriptions can be learned.
Everything here runs on one invented study — a six-week comparison of two statistics-course sections, phones parked in numbered pockets during class in one, used normally in the other, quizzed weekly. It is labeled illustrative wherever it appears. The research paper guide places both sections in the drafting order; this page zooms in on the line between them, including the merged-section case your brief may force on you.
The one-sentence difference
The discussion interprets; the conclusion resolves. That is the whole split. A sentence that argues — hedges a claim, compares your finding to prior work, weighs a rival explanation, admits a limit with what it threatens — does discussion work. A sentence that answers — resolves the opening question, folds findings into one pattern, lands settled advice and stops — does conclusion work.
Material travels between the sections: a limitation appears in both, an implication appears in both, and the section is decided by what the sentence does with it. The discussion unpacks the limit and ties it to a specific estimate; the conclusion names it once and moves on. Same fact, different job — the outline guide budgets the two sections separately.
The Section Splitter: ten sentences, one boundary
Each round shows one sentence from the illustrative study. Decide whether it belongs in the discussion or the conclusion and pick a side — you get the verdict instantly, plus a why-note naming the cue that gives it away. The score counter and progress meter fill as you go, a band-labelled result card lands at the end, and Reveal all answers shows the full deck with notes. Your score saves in this browser and returns on reload.
Section Splitter
10 sentences · a why-note after every pick
The sentence deck loads here — the splitter runs in your browser.
0/10
Clear split
What the discussion does: interpret, compare, limit, hedge
Four jobs, usually in this order. Interpret: say what the headline finding means and name its mechanism — the phone-pocket discussion argues that divided attention, not the phones themselves, did the damage. Results reported the rise; the discussion explains it. The full five-move structure lives in the discussion section guide.
Compare: put the finding next to the field. This is the only section where new citations belong, because comparison is the job — the discussion notes that earlier attention studies ran in labs while this one ran in a real classroom. If the comparisons feel messy, the literature review matrix template turns your sources into a liftable table.
Limit: name the two or three weak spots that actually touch your claims, each tied to what it threatens — "one course, six weeks" bounds the claim; a full term with finals might tell a different story. Hedge: calibrate every verb to the design — a two-section pilot can suggest, indicate and point; it cannot prove, demonstrate or demand.
What the conclusion does: answer, synthesize, land, stop
Answer first, in one breath. Copy the research question from your introduction and answer it in the same sentence — "this study set out to ask whether parking phones during class lifts quiz performance; the answer is yes." No warm-up before it, no new argument after it. The six-sentence loop shape lives in the research paper conclusion guide.
Synthesize: take the two or three headline findings and say what they add up to that neither says alone — "taken together, the six weeks of scores show that attention, not ability, was the bottleneck." Land: state who should act and how, as a settled recommendation — numbered pockets by the door costs an instructor nothing. A brief forward point is welcome: the next study, one changed element.
What the conclusion never carries: new citations, re-reported numbers, re-opened rival explanations. The abstract is the only section after it — the abstract guide shows the same study compressed for strangers — and it is written last, from these two sections.
The overlap zone: why the two sections blur
The boundary blurs because the sections share material, not because students are careless. Both sit at the end, describe the same findings, may name a limitation and an implication — and short papers often merge them on purpose, which makes the merge look licensed even when the brief asked for two. The failure mode is specific: you finish the discussion, open the conclusion, and copy the sentences that already worked — sixty seconds later it is a second discussion.
Journals merge them too — word limits make two separate sections a luxury in short empirical papers, and some venues use one combined heading. Merging is fine when it is chosen and signposted; it is a mistake when it is accidental. The test is never where a sentence sits but what it does — compare the two paragraphs below and watch the register change.
Discussion: The 1.4-point rise in the phone-parked section suggests that divided attention, not the phones themselves, did the damage. The pattern lines up with earlier attention studies, though most of those ran in labs rather than real classrooms. Because the phone-parked section met in the morning, the design cannot separate phones from time of day.
Conclusion: This study set out to ask whether parking phones during class lifts quiz performance; the answer is yes. The advantage held in every weekly quiz, and for instructors the step costs nothing: numbered pockets by the door and one standing rule.
Why it works: discussion sentences hedge, compare or limit; conclusion sentences resolve. Nothing is said twice — the second section compresses.
In conclusion, the results showed the phone-parked section scored higher. This suggests divided attention hurts learning, unlike previous studies which used self-reports, and future research should compare essay-based courses. The scores rose by 1.4 points on average, which proves phones should be parked, though one course may not generalize and more research is needed.
The diagnosis: one paragraph doing both jobs badly — a re-reported number, a verb that outran the design ("proves"), a discussion-side comparison, a rival explanation reopened at the closing bell, and "more research is needed" instead of a named next study. Split it at the last hedge: everything before interprets, everything after resolves.
The five-row differences table
When you are stuck on a sentence, find its row and follow it.
| Discussion | Conclusion | |
|---|---|---|
| Job | Interpret the findings and argue what they mean | Resolve the research question and close the paper |
| Tense | Past for the findings, present for the meaning — "this suggests" | Past for what was done, present for the settled answer |
| Citations allowed | Yes — comparison with prior work is the job | None new; nothing the paper has not already argued |
| Length | Roughly a fifth of the paper, second only to results | About five to ten percent — a tight paragraph or two |
| Hedging | Constant — every verb calibrated to the design | Minimal — the answer is stated as settled |
Run it as an audit on your own draft: underline every citation in the conclusion (row three allows none) and circle every hedge past the section break (row five spends them there). A draft that survives all five rows passes the split.
How graders split the marks between the two
Rubrics differ by department, so treat this as pattern-spotting, not a standard. On many empirical-paper rubrics the discussion is the heaviest or second-heaviest section — marks hang on interpretation, comparison with the literature, and limits tied to consequences — while the conclusion carries a smaller block earned by answering the question and stopping cleanly. Check your own brief: the weights decide where the effort goes.
The penalty patterns repeat more than the weights do. A discussion that re-reports results bleeds its own analysis marks — the grader wants meaning and finds measurement. A conclusion that re-argues gets capped: nothing new is claimed, so the extra words read as filler. A missing conclusion costs more than its mark count, because the introduction made a promise the paper visibly never kept. Every failure is one section doing the other's job.
When your paper has only one section
If the brief says "discussion and conclusion", or the word limit rules out two headings, merge on purpose. Run the discussion moves at full strength — interpret, compare, limit, hedge — then end the final paragraph on the closing beat a standalone conclusion would have carried: the answer, one synthesis sentence, one implication. A signpost — "what this study settles" — tells the reader the register changed.
Two details keep a merged section honest. First, keep the hedging in the front half and the resolution in the back half — a merged section that hedges its way to the last line never lands the answer. Second, resist a citation in the final paragraph; the combined heading does not relax the no-new-sources rule. The moves are the ones the discussion guide assembles; the closing beat is the loop the conclusion guide dissects.
Sentence-starter bank, split by section
Starters are scaffolding, not sentences to submit — replace every bracket, then check the claim still matches your design. If the claim feels wobbly, fix it first in the thesis statement guide; no starter survives a claim the evidence cannot carry.
- Weigh: The sharpest pattern in the results is [pattern], which points to [mechanism] rather than [rival explanation].
- Compare: This outcome lines up with [prior work] on [topic], and extends it by [the new wrinkle].
- Limit: [One course and six weeks] bound every claim, so [the estimate] should be read as [what it can support].
- Implicate: For [the actor], the finding justifies [the action] now, because [the cost or ease argument].
- Project: A follow-up across [settings], half [design A] and half [design B], would separate the two effects this study tangled.
- Answer: This study set out to ask [the research question]; the answer is [answer].
- Synthesize: Taken together, the findings show that [one pattern] — and it held across every test.
- Resolve: The advantage held in every weekly quiz, closing the question the introduction opened.
- Land: For [the actor] planning [the decision], the practical step is [the action].
- Hand off: The next study should test [one changed element]; this one has said what it can.
Three before-and-after rewrites
All three fixes come from the illustrative study. Read each "before", name what is wrong, then compare — the pattern behind every fix is the two-move test from the overlap zone.
Conclusion: This study set out to ask whether parking phones during class lifts quiz performance; the answer is yes.
Conclusion: For instructors who want to try this, the step costs nothing: numbered pockets by the door and one standing rule.
Conclusion: The question this study posed is answered, and the answer held all term.
Five common mistakes
The conclusion that re-opens the case
New interpretations, rival explanations or citations after the loop closed. Apply the two-move test: anything that argues moves up; what remains resolves.
The section break in the wrong place
Two headings, one blob: the break lands mid-argument, so the "conclusion" starts by hedging. Put the break where the last hedge sits.
Numbers smuggled into the conclusion
Re-reported statistics turn the final paragraph into a mini results section. Carry the direction of the effect, not the digits.
Apology inflation at the very end
Three hedged apologies as the last note read as panic. Unpack limits in the discussion; in the conclusion, name each big one once.
Verbatim echo pairs
The conclusion restates discussion sentences word for word. Compress instead — same claim, fewer words, new sentence — so the grader sees a paper that traveled.
Can your draft pass the split?
Tap each line as it clears. The meter tracks this visit only and resets when you close the tab.
0 of 6 checks
- Every interpretation, comparison and hedge sits in the discussion; the conclusion re-opens none
- The conclusion answers the research question in its first breath
- No number appears in the conclusion that results did not carry
- New citations live in the discussion only
- Each limitation is unpacked in the discussion and named once in the conclusion
- Read back to back, the two sections never repeat a sentence
Two sections planned, one deadline left?
When the split is clear but the hours will not stretch, a quoted, deadline-bound writer can draft both sections to your outline — and return them in time for you to defend every claim.
Five-question self-check
Five questions on the boundary — cues, citations, merged headings and the closing beat. Pick an answer and the question locks; wrong picks reveal the right one in place. Every answer appears on this page.
0 of 5 answered
1. Which sentence belongs in the conclusion rather than the discussion?
2. You find a new, relevant source while writing the conclusion. Where does it go?
3. Your brief asks for one combined "Discussion and conclusion" heading. What changes?
4. Where do unpacked limitations — limit plus consequence — belong?
5. Which closing line keeps the split clean?
Frequently asked questions
Where do limitations go — discussion or conclusion?
Both, at different depths. The discussion unpacks: name the limit, then say what it threatens — one course and six weeks bound every claim here. The conclusion names each big one once, plainly. If your limitation paragraph is longer than your synthesis sentence, the sections have swapped jobs.
Do I repeat my results in the conclusion?
No — you fold them. A conclusion that re-reports numbers is a second results section, and graders mark the redundancy. Take the headline findings and say what they add up to together: taken together, the six weeks of scores show attention was the bottleneck. Numbers stay in results; the direction travels.
Which section comes first — discussion or conclusion?
The discussion always comes first. The conclusion consumes it: it compresses the interpretation you just finished into a settled answer. Draft the discussion while the arguments are fresh, then close — and write the abstract last, once both exist. The outline guide budgets all three in the drafting order.
What heading do I use for a combined discussion and conclusion?
Whatever your brief or target journal specifies — one combined heading is common and acceptable. Inside, run the full discussion moves, then end on the closing beat a standalone conclusion would have carried: the answer, one synthesis sentence, one implication. A signpost tells the reader the register changed.
How do I fix a conclusion that reads like a second discussion?
Apply the two-move test. Every sentence either argues — hedges, compares, weighs rivals — or resolves. Move the arguing sentences into the discussion, and let what remains answer the research question in one breath. Most second-discussion conclusions shrink by half and get stronger.
Did the splitter make the boundary click?