Did you know that a classroom presentation can contradict itself even when its slide looks professional and its speaker sounds confident? Picture a pupil pointing at a graph showing a change from 18 to 15 litres, while announcing, ‘Our project reduced use by 30 per cent!’ A classmate notices that the numbers do not agree. That moment is not a reason for embarrassment; it is an invitation to practise evidence-led communication.
Resolving a disagreement between a presentation slide and a spoken claim means identifying the two different statements, checking the original data and calculation, correcting the unsupported version and telling the audience what changed. The immediate answer is not to believe the slide automatically or the speaker automatically. Both are representations made by the presenter; neither substitutes for the underlying evidence.
The diagnostic question: which statement is actually supported?
Before teaching any formula, show students a simple fictional slide: ‘Average recorded water use per pupil per day: before 18 L; after 15 L.’ Read aloud a fictional speaker’s sentence: ‘We reduced water use by thirty per cent.’ Ask learners to say what is different between the numerical display and the narration. A pupil who spots the conflict has taken the first step. The next test is whether they can calculate the percentage using the correct baseline and explain what the data do—and do not—establish.
The more subtle diagnostic is to swap the roles: give a slide with a mistaken figure but a correctly spoken correction. Can the student still check the underlying data rather than preferring whatever is written? A confident voice, a colourful chart and a neat caption are all presentation features. Evidence comes from the observations and the transparent method by which conclusions were drawn.
A worked example: water-use figures in a fictional school project
Every measurement and character in this section is invented for teaching. These are not real pupils’ results or a verified school water-saving outcome. Suppose a group has a comparable recorded daily average of 18 litres per pupil in period A and 15 litres per pupil in period B. Both values are labelled as litres per pupil per day. The group is describing a change between these periods, not a proven causal effect of its campaign.
First calculate the absolute change: 15 L − 18 L = −3 litres per pupil per day. The quantity fell by 3 L. For percentage change relative to the initial period, divide the 3 L fall by the initial 18 L: 3 ÷ 18 × 100% = 16.666…%. Round suitably to about 16.7% lower than the baseline. Check: 18 × 0.16666… = 3, so the computation is internally consistent.
A 30% fall from 18 L would be 18 × 0.30 = 5.4 L, giving a later value of 18 − 5.4 = 12.6 L, not 15 L. This countercheck exposes the mismatch immediately. It is impossible for both a change from 18 L to 15 L and a 30% fall relative to 18 L to be true under the same definition and rounding. The corrected spoken claim should be, ‘The two recorded periods differ by 3 litres per pupil per day, approximately 16.7% of the first period’s value.’
Do not jump from that arithmetic to ‘Our project caused the improvement.’ Other conditions may have changed: the number of days, weather, attendance, reporting rules or use of water outside the measured system. A comparison supports a measured difference if its data are credible; it does not by itself isolate the cause. That distinction between measurement, calculation and causal explanation is a central educational skill.
Three layers that audiences often confuse
- Primary record: the actual observations, collection dates, units, sampling and calculation method; this is where a numerical claim should be checked.
- Slide: a selected visual summary of the record, possibly simplified, rounded, mislabelled or edited.
- Spoken narration: the presenter’s explanation, which can add context but can also contain a slip, exaggeration or outdated figure.
These three layers are not three independent sources. A presenter may have copied one spreadsheet into a slide and then paraphrased the slide aloud. If both contain the same mistake, their agreement does not validate the claim. If they disagree, the group should return to the underlying observation and decide which representation needs repair.
A second example: words that imply a different proportion
Imagine a fictional class survey with 240 valid responses. A slide states ‘60% chose option A’, which corresponds to 0.60 × 240 = 144 responses, assuming the percentage is exact. The speaker says, ‘Three quarters of our respondents chose A.’ Three quarters means 75%, corresponding to 0.75 × 240 = 180 responses. Those claims differ by 36 responses, not just a stylistic nuance.
If the actual verified count was 144, the appropriate narration is ‘Three fifths, or sixty per cent, chose option A’—preferably with the count and denominator visible. If the raw tally was 180, the slide’s 60% would need correction instead. Neither percentage should be selected merely because it sounds more persuasive. Always confirm the survey denominator, whether nonresponses were excluded and how the options were defined.
An in-the-moment correction that preserves trust
Suppose a student notices the contradiction while presenting. A useful response is short: ‘Correction: I said thirty per cent, but the figures shown are 18 and 15 litres. The reduction between those recorded averages is about 16.7 per cent. We will correct the speaker notes after the presentation.’ This does not erase the slip, but it avoids leaving the audience with two incompatible figures.
If the student is unsure which version is supported, the better statement is: ‘Our slide and narration disagree; we need to check the original table before asserting a percentage.’ That is preferable to guessing in front of an audience. A teacher can make room for accurate correction without rewarding unearned certainty or punishing every honest pause.
For a published slide deck or shared handout, label the corrected version with a date or version note, and retain the original data and calculation. A useful correction says what was wrong and what the corrected value is; it does not quietly replace a figure and pretend no discrepancy occurred. This is especially important if an audience later cites the slide.
How to build a claim-to-evidence cross-check
- Write the claim: copy the exact words spoken, including qualifiers such as ‘about’, ‘all’, ‘average’ or ’caused’.
- Find the display: note the slide number, graph axis titles, units, time period and any denominator.
- Locate the source: identify the original table, worksheet, notebook entry or other verifiable record behind the display.
- Recompute independently: redo the arithmetic using the intended formula and check whether rounding could explain a small discrepancy.
- Repair the communication: correct the slide, the script, or both; state the limit of what the evidence proves.
- Retest with a reader: ask a classmate to read the slide and paraphrase the intended claim without hearing the speaker.
An effective team assigns a checker to the number, a checker to the language and a presenter to the explanation. The point is not to create a bureaucracy for a small project. It is to prevent one unverified sentence from travelling unchanged through slides, speech, handouts and social posts.
Practice that progresses beyond finding an obvious error
Guided: spot the mismatch
Provide a fictional slide with 40 books borrowed in week one and 50 in week two. The speaker says ‘book borrowing doubled’. Learners calculate the change: 50 − 40 = 10 books; relative increase is 10 ÷ 40 × 100% = 25%. Doubling would require 80 books in week two. Rewrite the claim: ‘Borrowing rose by 10 books, or 25%, between the two weeks.’ Explain that the counts alone do not prove why borrowing changed.
Supported: decide whether the slide or narration is wrong
Use a fictional project table containing 12, 15 and 18 minutes for three tasks. Their sum is 45 minutes. Show a slide reading 45 minutes but narrate ’55 minutes’. Have learners recount the three values, then correct the spoken total. Swap the example: show a slide reading 50 minutes and narrate the correct 45. This prevents the habit of automatically trusting visual text over speech.
Independent: audit a full three-slide story
Ask each group to make a short educational presentation with one numerical observation, one interpretation and one limitation. Another group must compare the speech script, the slide and the supplied source table, mark every disagreement and propose a correction. A high-quality answer includes units and denominators, catches unsupported cause-and-effect language and names at least one thing that cannot be inferred.
For a stronger transfer task, give students a bar chart in which the scale starts above zero or an infographic that displays a percentage without a sample size. The presenter’s spoken claim might not contain an arithmetic mistake but could still exaggerate what the display demonstrates. The audience must scrutinise the communication and the evidence together.
What teachers and parents can measure
A teacher can collect a before-and-after sample of a learner’s script and ask: Did the student correct the numerical claim? Did the revision preserve the unit and denominator? Did the final sentence express a supported conclusion rather than a stronger promise? The improvement is observable in the text, even without assigning a score. It may be helpful to have students explain a corrected slide in their own words rather than merely read a prepared phrase.
Parents can rehearse a presentation by listening once without looking at the slides, then checking whether the picture supports what they heard. Ask, ‘Where did this number come from?’ and ‘What would make this conclusion less certain?’ Avoid rewriting every child’s words for them. The goal is for a learner to notice discrepancies and self-correct in an unfamiliar topic.
Limitations: numbers, uncertainty and audience questions
Not every mismatch signals an error. A speaker may intentionally use a rounded number, a different denominator or a different time range—but those differences should be labelled. For example, 16.7% and ‘about 17%’ agree under ordinary rounding; 16.7% and 30% do not. If a spoken number refers to a later updated dataset, the slide must indicate that the two datasets differ. Clear attribution prevents an innocent format difference from becoming a misleading assertion.
Some conclusions depend on measurements with uncertainty, small samples or incomplete records. Students should not use more decimal places than their data justify, claim precision they did not measure, or imply that a before-and-after comparison automatically proves causation. An audience question that exposes an uncertainty can be answered with a limitation, rather than an invented explanation.
Frequently asked questions
Should a pupil always trust the written slide over the speaker?
No. Both can be wrong. Compare the original recorded evidence, check the calculation and clarify whether the two claims describe the same quantity, group and period.
What if the figures differ only because of rounding?
Check the underlying exact values and the stated rounding convention. ‘16.7%’ and ‘about 17%’ are compatible; a major difference such as 16.7% versus 30% requires further explanation.
Does correcting a presentation make the learner look less confident?
A clear correction may demonstrate better judgement than defending an unsupported number. The educational goal is warranted confidence, not the appearance of infallibility.
Can a slide be accurate while its conclusion is too strong?
Yes. A graph may faithfully show two measured values while the narration asserts a cause that was not tested. Verify the transition from the data to the claim, not just the arithmetic.
Further learning and source guidance
Read how to check a school project conclusion against its own graph, then visit the eduKate Education Hub for broader independent-learning approaches. The Clementi small-group teaching reference illustrates diagnosis followed by guided practice and self-checking.
For presentation design, the MIT Communication Lab’s slideshow guide discusses making figures support the message, while its policy-presentation guide distinguishes what figures, tables and spoken text communicate. These are practical presentation resources; the fictional classroom figures above are our own teaching examples.
Learning transfer: a good presentation is not a competition between slides and speech. It is a consistent, checkable explanation in which the numbers, the language and the conclusion all point to the same evidence.