A lab report conclusion answers the question the introduction asked, in five moves: restate the main finding with its direction and units, say whether the data support the hypothesis, name the largest error source and what it would do to the result, state what the finding means, and propose the next experiment. Support — never prove — is the verb that keeps the section honest.
Conventions vary by course, and the lab manual and rubric decide the final word — section naming, tense exceptions and length all flex. This page covers the closing section only: the full IMRaD structure, the section-by-section planner and the voice drills live on the sibling guide, how to write a lab report. Every sample sentence here is an illustrative placeholder with [brackets] for your data — structure practice, never a real experiment.
What the conclusion of a lab report must contain
The closing section gets read twice: once by the grader deciding whether the report hangs together, and once by the grader who only has time for the abstract and the ending. Both readings ask the same question — did the experiment produce an answer, and does the writer know what that answer is worth? Five moves answer it. They run in order, and each has a failure mode graders know on sight:
| Move | Its job in one sentence | Failure mode graders flag |
|---|---|---|
| 1. Restate the answer | Open with the one-sentence answer to the question the introduction asked — with direction and units. | An activity summary ("we did X, then Y") instead of a finding. |
| 2. Compare with the hypothesis | Give the verdict: the data support, partly support or contradict the prediction — and say which part did what. | "It proved" — or no verdict at all, leaving the reader to guess. |
| 3. Analyze the error | Name the largest source of error and the direction it would push the result. | "Human error" as a catch-all, or a limitation list with no consequences attached. |
| 4. State the implication | Say what the finding means for the system studied, inside the limits of the design. | Grand claims the dataset cannot reach — the cure, the product, the policy. |
| 5. Propose future work | Hand over one concrete, doable follow-up experiment. | "More research is needed" — with no research named anywhere. |
The order does work that reordering destroys. Restating the answer first anchors everything after it; the hypothesis comparison turns the result into a verdict; error analysis prices the verdict; implications and future work spend it. A conclusion that opens with the follow-up proposal — the most common student ordering — makes the grader wait for the point, and graders do not wait.
Naming differs by course: some rubrics call this closing section the Discussion, some the Conclusion, and some ask for both. The five moves fit every version — a merged section runs them at full length, and a separate short conclusion runs them compressed. For the long-paper version of the same craft, the discussion section guide goes deeper on literature comparison, and the discussion vs conclusion breakdown maps where one section ends and the other begins. The whole-report craft — the IMRaD spine, section budgets, voice and tense drills — stays on the lab report guide, which this page does not repeat.
The conclusion builder
Pick two things — what the data did to your hypothesis, and what the data actually look like — and the builder assembles a five-move draft. Each move arrives as a starter sentence with [brackets] where your facts go: replace the brackets, keep the structure, and the draft becomes a conclusion. The refills respond to honest picks, because "partly supported" reads differently from "refuted", and "one odd run" writes a different error analysis than "no clear pattern". Your picks and text persist in this browser only — nothing is sent anywhere — and Reset wipes the lot.
Conclusion builder
100% private — runs in your browserOpen with the one-sentence answer to the question the introduction asked — direction and units included.
Give the verdict: support, partial support or contradiction — and which part did what.
Name the largest source of error and the direction it would push the result.
Say what the finding means for the system studied, inside the limits of the design.
Give one concrete, doable follow-up that follows from the result.
Changing a pick refills all five starters — so settle the picks first, then write your versions, or the refill will overwrite them. Every bracket is a slot for your data, and the draft is a structure to adapt, not prose to transplant.
Two rules keep the draft honest. First, the starters are scaffolding: your instructor has read these shapes before, so swap the wording once the structure is right. Second, the builder cannot know your data — every bracket is a claim about your experiment that only you can fill, and an unfilled bracket that reaches submission reads as exactly what it is.
Sentence starters for each move
The builder hands you one starter per move, but graders read stacks of conclusions, so it pays to own two or three ways into each one. The bank below gives alternates — paste them, or steal the shape. All are illustrative structures with [brackets]:
| Move | Alternate starters |
|---|---|
| Restate the answer | "The experiment set out to [aim] and produced a clear result: [answer with direction and units]." · "Measured against [the aim], the data answer the opening question: [answer]." · "The central result is [finding], and it answers [the question] directly." |
| Compare with the hypothesis | "The outcome matches the prediction that [prediction] for [component], though not for [component]." · "Taken together, the trials support a narrower version of the hypothesis: [narrowed claim]." · "The data contradict the prediction on [specific point], which suggests [the assumption that failed]." |
| Analyze the error | "The likeliest source of error is [source], expected to shift [measure] by [direction or amount]." · "Two limitations bound this result: [limitation A] and [limitation B]; the first is the larger." · "The deviation in trial [x] is consistent with [cause] rather than with a property of the system." |
| State the implication | "For [the system studied], the finding means [practical or conceptual consequence]." · "The result matters for [course topic] because [reason]." · "At minimum, the data justify [the modest claim the design supports]." |
| Propose future work | "A natural next step is [concrete change], which would decide between [explanation A] and [explanation B]." · "Repeating the trial with [one variable changed] would test the [error source] explanation directly." · "Extending [the range] to [value] would show whether the trend persists." |
Starter or no starter, the section rises or falls on verb choices. Three pairs below show the rewrite that turns a flagged sentence into a kept one:
Weak: "In this lab we measured [measure] three times, then changed [the factor] and measured it again."
Why it is weak: it narrates the session schedule, not the finding. The reader wants the answer in the first sentence — the procedure already has a whole Methods section to live in.
Strong: "Changing [the factor] raised [measure] from [value] to [value], answering the question the introduction posed."
Why it is strong: the finding leads, the direction and values carry it, and the loop back to the introduction closes the report as one piece of writing.
Weak: "The experiment proved our hypothesis was correct beyond any doubt."
Why it is weak: proof is a mathematical standard, not a measuring one, and "beyond any doubt" claims certainty the design never tested for. One dataset under one set of conditions cannot carry that weight, and graders mark the overclaim.
Strong: "The results are consistent with the hypothesis: [the specific evidence] matches the predicted [direction], within [the uncertainty]."
Why it is strong: support is the honest ceiling for measured evidence, and naming the uncertainty shows the verdict was priced, not declared.
Weak: "More research is needed to fully understand this topic in the future."
Why it is weak: it names no research, no change and no question — it is a throat-clear disguised as a plan, and it wastes the move the rubric set aside for follow-up thinking.
Strong: "A follow-up run with [the error source] controlled and [the range] extended to [value] would test whether the trend holds."
Why it is strong: it names the experiment, the variable it fixes and the question it answers — a grader could hand it to the next class as an assignment.
The banned-phrase fixer
Some sentences cost marks on sight — not because they get the science wrong, but because they promise certainty the design cannot deliver or skip the analysis the move requires. Six of the most common appear below, one at a time. Pick the fix you would actually submit, and the drill explains what changed. Nothing here is stored or scored anywhere; the phrases are illustrative.
Banned-phrase fixer
Six phrases · pick the honest fixLoading…
The pattern behind all six: certainty verbs, experience talk and category words — "human error", "more research" — stand in for the specific work the move asks for. A quick self-test before submission: if a draft sentence cannot name its number, its direction or its source, it is not finished yet.
Word budgets by report length
Conclusion length is a budget decision, and the budget comes from the report's total size. The split below is a structural starting point — the rubric overrides it — but it prevents the two classic shape failures: the two-line ending stapled to a full report, and the half-page epilogue stapled to a ten-minute lab.
| Report size | Total words | Closing section | How to spend them |
|---|---|---|---|
| Short write-up | 600–1,000 | 100–180 | One tight sentence per move; the error note can share its sentence with the verdict. |
| Standard report | 1,200–2,000 | 250–450 | All five moves at full length; error analysis takes the largest single block. |
| Extended report | 2,500+ | 500–900 | Discussion and conclusion may split into two sections; the literature comparison and error analysis grow, and future work earns a short paragraph. |
Where the rest of the words live is the other half of the budget: the general report guide shows how every section takes its share, and the report writing format page covers the front matter and heading anatomy around them. One special case: in a progress report, the future-work move doubles as the work plan for the next period, so it grows while the verdict move shrinks — the hypothesis there is still being tested, not yet settled.
Report written, conclusion stalled?
If the draft is done except for the closing section, a quoted, deadline-bound writer can shape the discussion and conclusion around your results while the experiment and the data stay yours. Get a free price quote and see the terms first.
Quiz: five questions on closing a report
Five questions on conclusion craft — the moves, the verbs, the budgets. Every wrong answer explains itself, and four options keep the guessing honest.
1. What is the first job of a lab report conclusion?
2. The hypothesis predicted a rise in [measure], but the values fell. The comparison move should say:
3. Which sentence does the error-analysis move actually require?
4. What is wrong with "It proved that [factor] doubles [measure]"?
5. Which future-work sentence earns its place?
Frequently asked questions
What should the conclusion of a lab report include?
Five moves, in order: restate the one-sentence answer to the question the introduction asked; say whether the data support, partly support or contradict the hypothesis; name the largest source of error and the direction it pushed the result; state what the finding means within the limits of the design; propose one concrete follow-up experiment. What stays out: new data, new analysis and claims the dataset cannot carry.
How long should a lab report conclusion be?
The rubric decides, and where it is silent the report's size sets the budget: roughly 100–180 words for a 600–1,000-word write-up, 250–450 for a standard 1,200–2,000-word report, and 500–900 when the discussion is a separate, formal section. Inside any size, error analysis earns the largest single share — a conclusion that never limits itself is the shortest route to lost marks.
Do you restate the hypothesis in the conclusion?
Yes — the comparison move is the whole point. State whether the results support the prediction, support part of it, or contradict it, and say which part did what. What you should not write is that the results proved it: a single dataset supports or is consistent with a hypothesis, and the proof verb overclaims in both directions.
Can you say the results proved the hypothesis?
Better not. One experiment produces one dataset under one set of conditions, which can suggest, indicate or be consistent with a hypothesis — but proving is what mathematics does, not measurement. Graders read it as the surest sign the writer cannot yet separate evidence from certainty, and mark the comparison move accordingly.
What is the difference between a discussion and a conclusion in a lab report?
In many courses the two words name one merged section, and the five moves cover it either way. Where they are separate, the discussion carries the comparison to literature and the full error analysis, while the conclusion is a tight restatement: answer, verdict on the hypothesis, and the next step. Your lab manual's section list outranks any convention on this page.
Did this page get your conclusion unstuck?