The Core Aim of Secondary 4 Vocabulary | Writing a Source-Limitation Note Before a Strong Claim Becomes an Overclaim

You are here because a Secondary 4 argument, report or oral explanation sounds convincing, yet its evidence supports something narrower than the student has claimed. One careful note about what a source can and cannot show can save the whole argument from overstating its case.

The practical answer is to put a four-part source-limitation note beside each consequential claim: what this source actually establishes; what it leaves unknown; how far the present claim may go; and what evidence would be needed to make a stronger claim. “Three interviewed pupils preferred the change” does not mean “the entire school supports the change”. The aim is not timid writing. It is confidence proportional to evidence.

Start with the answer you need, not the whole archive

If you have a persuasive essay due, begin with the claim-limit test and the model evidence ledger. If you have been told that your writing overgeneralises, go to the vocabulary boundary clinic. If you are evaluating an unfamiliar report, jump to the source casebook. If you are practising independently, use the explained-answer laboratory. The whole piece is deliberately long, but the first good repair should be available without reading every later case.

Why a source-limitation note is a literacy tool, not an apology

A limitation is a truthful statement about what a source cannot reliably establish. It is not a compulsory self-criticism placed at the end of an essay to sound academic. Suppose a student interviews five classmates about a proposed change to the school canteen. Four welcome it. The writer may legitimately report that four of five interviewed pupils welcomed the proposal. They may explore the reasons those pupils gave. They may not safely conclude that all Secondary 4 students—or all parents or teachers—support the proposal, because the interview did not sample those groups. That is a scope problem. A useful limitation note points directly to the missing reach of the evidence.

A strong argument does not depend on pretending every source proves everything. It becomes stronger when each source has a defined role. An official timetable can establish scheduled times. An eyewitness can describe what they personally noticed. A survey can report respondents’ answers, with sampling and question-wording limitations. A teacher’s classroom observation may support a claim about one group over a period, but not a universal claim about all schools. A source is useful because of its fit for the question, not simply because its title looks impressive.

The vocabulary in this article—indicates, suggests, documents, estimates, reports, observes, infers, samples, attributes, contradicts, qualifies—functions as the control panel of a claim. It tells a reader where the knowledge came from and how far they can travel with it. The aim is not to dilute every sentence with “perhaps”. A timetable that explicitly says the event begins at nine permits the firm sentence “The published timetable states a 9 am start,” while leaving space for last-minute changes. Saying “possibly the timetable might say…” would be needlessly evasive.

The source-limitation note is also a reading comprehension exercise. To write a limitation, the student must accurately read what the source claims, identify its population and time, notice what it does not mention, and distinguish the author’s inference from the original data. This pushes vocabulary beyond dictionary memorisation: the same words that make an argument honest help a student understand an article, assess a visual text, summarise without drifting, and defend a decision in spoken discussion.

For 2026 Singapore Secondary 4 learners, exam and school contexts vary by subject level and programme. The SEAB 2026 GCE O-Level school-candidate syllabus listing includes English Language 1184. This article is an original teaching resource, not a prescribed SEAB task, an official examination marking guide or a claim that every Secondary 4 pupil follows the same qualification route. From 2027, the Singapore-Cambridge Secondary Education Certificate changes the examination context for eligible cohorts, so readers should check the relevant year rather than silently treating different exams as identical.

Diagnose the leap: locate exactly where evidence turns into certainty

Begin with a sentence, not a general judgement about the writer’s character. “All teenagers learn better from video” is an unusually broad claim. Before searching for a substitute adjective, ask five questions: Who is included in “all”? What is meant by “learn better”? What kind of video? Compared with what method? Under what conditions? A source showing higher average recall for one small group after one short lesson might be relevant to the question, yet still cannot establish an advantage for every adolescent and subject. The gap lies between the source’s design and the sentence’s reach.

A simple diagnostic method is to underline quantifiers in red—all, every, never, always, most, many, some, one—and cause words in blue—because, caused, prevented, made, led to, resulted in. Then circle time markers and group names. Those are the places where claims often expand beyond evidence. Ask the student to restate the actual source in one sentence without persuasion. If the sentence remains broader than the source, the first repair is comprehension and scope, not style.

Next, distinguish four common failures. A population leap moves from a few observed people to a much larger group. A time leap moves from an old observation to an unstated current claim. A causal leap turns association or sequence into a demonstrated mechanism. A definition leap changes what a key term means between the source and the conclusion. Each failure demands a different repair, and the words chosen should reveal which boundary was crossed.

A student might avoid all four leaps and still misuse quotation. Suppose a writer condenses an interviewee’s long reply into six polished words inside quotation marks. The message may approximate the speaker’s view, but it no longer necessarily represents the exact wording. The repair is to quote accurately in context or to paraphrase with attribution. Another common problem is what the source omits: a photograph records one selected instant and frame, while a written report may supply an extended sequence. The writer cannot invent the absent information because it would make the conclusion more interesting.

A diagnosis can be more than one word, but should remain small enough to guide teaching. “This sentence claims all commuters but the survey covers twenty volunteers at one station on a weekday morning” is better than “biased”. “The article reports that the number increased after the new rule but provides no comparison group, so it cannot isolate that rule as the cause” is better than “unreliable”. The specific relation matters. A source can be valuable for one question and unsuitable for another without becoming worthless.

Use the four-column LIMIT note

LIMIT here is a learning mnemonic, not an official research standard. The first column names the literal evidence. The second identifies the included and excluded scope. The third proposes the maximum defensible sentence. The fourth states the additional test or source needed. Students should use the four fields to reason, not to fill space. If a clear two-sentence note does the job, there is no need to inflate it.

Literal evidence	Scope and limitation	Defensible claim	Needed next evidence
Four of five interviewed classmates favoured the canteen change	Small voluntary set; question wording and views elsewhere unknown	Four of the five classmates we interviewed favoured the change	Wider, well-described sample and consistent questions to discuss school-wide patterns
A display notice says the rehearsal begins at 4 pm	Published schedule at a given time; changes not checked	The posted notice lists a 4 pm start	Latest authorised update for a claim about the actual start
A fictional diary describes one commuter’s difficult journey	One personal account; cannot establish typical commute	The writer describes that particular trip as difficult	Comparable accounts or wider transport data for frequency claims
A graph rises after a new school routine begins	Temporal pattern; confounding variables not resolved	Recorded values rose during the period following the change	A credible comparison and explanatory method to assess causes

Read the second row carefully: a notice is appropriate evidence of what was scheduled. It is not proof that an event actually started on time. A later eyewitness account might address the latter question; even then, a single recollection may need corroboration if the exact minute matters. This is not a cynical refusal to trust notices or people. It is matching a source’s properties to the claim the writer wants readers to accept.

The fourth column should never become a fantasy demand for impossible certainty. Sometimes further evidence cannot reasonably be gathered. A student can then state a modest supported conclusion and its limitation. “We cannot determine whether the rule caused the change from these figures alone” is a useful finding. It prevents a leap while keeping the observed increase available for future discussion. Not every school essay needs to establish causal truth at research level; the point is to represent uncertainty honestly within the task’s expectations.

Boundary words: choose a claim strength that the text earns

Reports, shows, suggests and proves

Reports tells the reader that a source states something. It does not automatically endorse the statement as true. “The witness reported hearing a bell” is faithful if the witness gave that account. Shows can be appropriate for directly represented data—“the chart shows three recorded years”—but may become too strong when applied to a causal conclusion. Suggests marks a reasonable but incomplete inference. Proves carries an especially heavy burden and is rarely justified by a single classroom example or small observational dataset. The right choice depends on evidence, not how certain the writer wants to sound.

Try a substitution test. “The school survey proves pupils prefer later starts” becomes “In this voluntary survey, a majority of respondents said they preferred a later start”. The second version may sound less dramatic, yet gives a reader more information: respondents, scope, instrument and the reported preference. If the survey did not disclose how participants were recruited, even “majority” should remain tied explicitly to those who responded, not the whole school.

Caused, contributed to, coincided with and followed

A claim of cause identifies a mechanism or credible design that connects an event to an outcome. Contributed to is not a free shortcut: it still asserts some causal role. Coincided with reports temporal overlap. Followed establishes sequence. These words are often treated as interchangeable connectors in weak essays, but they lead readers to different conclusions. When a graph rises during a campaign, “coincided with” may be safer than “caused” unless comparisons and methods support a causal inference.

Yet over-caution also distorts. If a simple fictional text explicitly says a ball struck a window and shattered it, “The impact shattered the window” is a supported narrative fact. A limitation note should not insist that a reader cannot know causation in every ordinary story. The skill is to distinguish what the given source explicitly establishes from what its design leaves open. Excessive qualifiers can make a clear account evasive; missing qualifiers can make a weak inference seem settled.

All, most, some and observed participants

Quantifiers are easy to notice in isolation and remarkably easy to mishandle when persuasive pressure rises. “All” usually requires a universal reference group and no known exceptions. “Most” generally means more than half of a specified group, but a survey’s respondents are not the same thing as every member of that population. “Some” indicates a nonzero subset; its exact proportion may be unknown. Often the most accurate formulation simply names the observed count and group: “Seven of the ten pupils who completed the task…”

A sentence such as “Teenagers disagree with adults” hides variation and lacks a reference context. An observation about three pupils disagreeing during a class debate can be described faithfully without becoming a permanent claim about a generation. Specific groups may still show important patterns, but evidence about those patterns must be appropriate. This is where lexical precision supports fairness: the writer does not erase differences among people merely to make the conclusion memorable.

Current, at the time, estimated and projected

When a source contains a date, carry that date into claims that depend on it. A timetable published in 2023 may no longer describe a 2026 route. A population estimate refers to a measurement or estimation period; a projection concerns a possible future based on assumptions. “Current” should have a reasonably recent check and a clear reference point in time-sensitive contexts. Recording the access date helps but does not transform an obsolete source into current evidence.

In a revision exercise, students should ask two questions: Was this the source’s latest version when the claim was written? And does the conclusion need to be true now, or merely true of that documented historical period? A history assignment can correctly cite an old newspaper to show contemporary opinion at the time. The same newspaper may be unsuitable as a factual guide to present-day transport fares. Source age is not a universal quality score; relevance to the reader’s question is the real test.

Why limitation vocabulary is useful across English papers and beyond

For school reading comprehension, a source-limitation note trains the habit of not importing an unsupported detail into an answer. For situational writing, it stops a student promising a change or outcome that the prompt does not authorise. For continuous writing, it helps an argument distinguish personal experience from general evidence and helps narrative commentary avoid inventing what a character knows. In spoken interaction, it offers short, fair phrases for qualified agreement and disagreement. The same principle travels to Science when a diagram or experiment cannot establish everything; to Humanities when sources differ; and to everyday decisions when statistics appear in an advertisement.

The connection to O-Level English should be accurate rather than grandiose. SEAB’s 2026 English Language syllabus 1184 has Writing, Comprehension, Listening and Oral Communication components. Its situational writing assesses purpose, audience and context, while continuous writing can involve an argument. This article does not claim that students will be asked to submit a “LIMIT note”, nor that formal source evaluation earns guaranteed marks. The note is a rehearsal strategy that makes evidence and claims easier to check before they enter the actual answer.

A productive Secondary 4 learner can begin to decide which vocabulary matters for a given reader. Sometimes they should use a simple sentence: “We do not know how the participants were chosen.” Sometimes they may write a more qualified explanation about sample bias, time scope, confounding factors or document provenance. The test is not the density of difficult nouns. It is whether another person can reconstruct how the conclusion follows—or does not follow—from the source.

Two model revisions from claim to limitation and back

Model 1 — a school rule and attendance

An invented report states: “After the school introduced quiet reading in the first ten minutes of registration, the proportion of late arrivals fell from fourteen to nine per hundred recorded sessions.” A student’s original argument reads, “Quiet reading solved lateness.” This is persuasive but goes further than the record. We do not know whether transport arrangements, timetable changes, weather, attendance policies or recording methods also changed. We also do not know how often or for whom late arrival was measured. The correct first move is not to throw out the observed pattern; it is to keep it with the proper scope.

A careful replacement is: “The school’s recorded late-arrival proportion fell during the period after quiet reading was introduced. This timing is compatible with the routine helping, but the figures alone do not establish that the routine caused the decrease.” The limitation note says what was counted and what remains untested. An additional comparison across relevant periods or groups, with consistent measurement and attention to other changes, would strengthen a causal evaluation. A good paragraph can still recommend trying the routine; it should distinguish a practical proposal from an established explanation.

Model 2 — a striking individual quotation

An invented class interview records one pupil saying, “I can concentrate much better when everyone leaves their phones outside.” A student’s argument concludes, “Phone restrictions improve all students’ concentration.” The leap crosses from one subjective report to a universal intervention outcome. It also equates feeling more focused with a directly measured performance gain. Those distinctions matter because a quotation can be vivid and relevant without being a representative study.

A more faithful version is: “One pupil reported finding it easier to concentrate when phones were kept outside the room. This account shows a perceived benefit for that pupil, but it does not establish the effect on every student or whether their work improved.” A writer could pair this with other appropriate sources to discuss the possible merits and costs of a phone policy. A balanced argument is not obliged to agree and disagree equally with every position; it is obliged not to hide the limits of the support it uses.

Claim calibration as a two-minute independent check

• Circle the claim word carrying the widest reach: every, always, caused, entirely, proves, typical, current.
• Name the source type: interview, timetable, data table, image, press release, historical diary, survey, experiment or secondary summary.
• Write the narrowest faithful summary in one sentence before attempting persuasion.
• Test population, time, place, measurement, comparison, quotation, cause and missing context.
• Revise the original claim or acquire better evidence. Do not preserve the stronger wording simply because it sounds impressive.

The test should remain proportionate. A claim that a fictional character entered a room may need only a clear line from the story. A claim that an intervention works for every child, or a factual allegation about a real person, demands far more care. Students should learn to increase checking in proportion to the importance of the claim. One useful sign of maturity is being able to say “I have enough evidence to make this smaller statement” rather than using uncertainty as a reason to make no statement at all.

Source casebook A — What the evidence actually records

These sixteen fictional case files develop the first source-limitation habit. Each begins with a tempting inference, identifies the exact point at which it goes beyond the evidence, and supplies a revised statement. The practice questions are intentionally different from the model. Answers are explained so readers can check their reasoning rather than copy a phrase and mistake familiarity for mastery. None of these invented records represents a real school, pupil or research outcome.

1. One photograph made to describe a whole event

A student sees a fictional photo showing a crowded bus platform at 7.45 am. The accompanying argument says that the station is overcrowded throughout the entire day. The photo has a timestamp but no sequence of images or passenger counts.

Diagnose the claim leap: The source captures one viewpoint at one instant. The writer has expanded a moment into a duration and a visible arrangement into a measured capacity problem. A busy platform may warrant further investigation without proving conditions at every hour.

Supported revision: “The photograph shows a crowded-looking platform at 7.45 am on the day recorded. It does not establish whether crowding continued throughout the day or exceeded an operational threshold.”

Unseen practice: An image shows empty seats inside a library at closing time. Can it prove the library is unpopular? Write a limitation note with one possible next source.

Explained answer 1

No. It shows seats empty at that time and from that camera angle. Attendance records covering different hours, alongside relevant context, could address broader use; one closing-time image is not a reliable popularity measure.

Changed condition: A second image shows the same platform empty at noon. The conclusion should now acknowledge variation across two times rather than choosing whichever image suits the argument.

What this teaches: Photographic evidence can be real and still have a narrow time and framing boundary.

2. An eyewitness’s interpretation made into an observed fact

After a school play, a pupil tells a reporter that a stage manager “looked annoyed”. The draft account claims that the manager was angry with a particular performer and planned to remove them from the cast.

Diagnose the claim leap: The eyewitness described a visible impression. The written conclusion adds an emotion’s exact target and an intended action that were not reported. Attribution and uncertainty have disappeared together.

Supported revision: “One attendee described the manager as looking annoyed. The available account does not establish the manager’s feelings, their cause or any decision about casting.”

Unseen practice: A spectator says a judge “seemed disappointed” after a result. Write one faithful sentence and one statement that would need independent support.

Explained answer 2

“A spectator thought the judge seemed disappointed” is attributed. “The judge decided to penalise the team out of anger” would need direct evidence of the decision and its reason.

Changed condition: If a public statement later explains that the manager was concerned about a safety procedure, the report can quote or paraphrase that explanation with a source tag. Do not silently rewrite what the eyewitness originally saw.

What this teaches: An observer can report behaviour; motive is a separate evidential question.

3. A quote shortened until its condition disappears

An invented interviewee says, “I support using the hall if the organisers can keep the side entrance clear.” The student quotes only, “I support using the hall,” and describes unconditional approval.

Diagnose the claim leap: A conditional statement has been converted into a general endorsement. The missing if-clause could materially alter a reader’s decision. Ellipsis is not an excuse to cut context that changes meaning.

Supported revision: “The interviewee supported using the hall subject to keeping the side entrance clear.” If using direct quotation, retain the relevant condition in the wording and make any editorial omissions transparent.

Unseen practice: The original answer says, “I may join the trip if it ends before six.” A summary reads, “She will join the trip.” Repair it.

Explained answer 3

“She said she might join if the trip ended before six.” The revision preserves both possibility and condition. “Will” wrongly converts interest into a firm commitment.

Changed condition: The event timetable later confirms a 5.30 pm finish. The condition is now satisfied, but the person still has not necessarily confirmed attendance; a second response may be needed.

What this teaches: A limitation note protects the condition that gives a quotation its meaning.

4. A press release treated as independent evaluation

A fictional technology supplier issues a press release saying its new classroom timer “transforms attention”. A student writes that research has proven the timer improves learning. The release describes the supplier’s own promotional claims but no independent study.

Diagnose the claim leap: The document exists and is relevant as evidence of what the company claims. Its purpose and missing method limit its role as evidence that the product works. Calling it false without checking would be another overclaim.

Supported revision: “The supplier claims that the timer improves attention. The release does not supply enough independent evidence here to establish how much it changes learning outcomes.”

Unseen practice: A school-café vendor advertises that its snacks are “the healthiest choice”. What is the strongest defensible claim from the advertisement alone?

Explained answer 4

“The vendor advertises the snacks as a healthy choice.” Nutritional comparison would require ingredients, an appropriate health criterion and a relevant assessment rather than merely the vendor’s slogan.

Changed condition: If an independent review later publishes a specified outcome under described conditions, incorporate that evidence and distinguish the tested outcome from the supplier’s broader slogan.

What this teaches: A source can reliably document a speaker’s assertion even when it cannot establish that assertion’s truth.

5. A timetable read as evidence of what happened

A notice lists a performance scheduled to begin at four. A student writes that the performance actually began at four and finished on time. No attendance, recording or contemporaneous event report was supplied.

Diagnose the claim leap: Scheduled and actual are different statuses. The notice may be authoritative about the planned start; it cannot by itself verify real execution or the ending.

Supported revision: “The published notice scheduled the performance for 4 pm. The available document does not confirm the actual starting or finishing time.”

Unseen practice: An online event page lists the debate at 3 pm; a student’s account says “The debate lasted one hour”. Which part lacks evidence and what should be checked?

Explained answer 5

The advertised start may be sourced to the page. The duration needs an actual programme, recording, organisers’ account or similarly relevant evidence. Do not infer duration from the timetable entry.

Changed condition: After a recording confirms that the debate began at 3.12 pm, revise the actual start while retaining the scheduled 3 pm time as historical context when useful.

What this teaches: An official plan can be dependable without proving every subsequent event matched it.

6. A forecast recast as an accomplished outcome

A fictional council report projects that a walkway may reduce crowding by 15 per cent under a particular model. A student’s speech states that the walkway has already cut congestion by 15 per cent, even though construction has not begun.

Diagnose the claim leap: A modelled future possibility is neither a measurement of realised effect nor a guaranteed consequence. The projected figure depends on assumptions that a responsible summary must not erase.

Supported revision: “The council’s model projects a possible 15 per cent reduction under its stated assumptions. The walkway’s actual effect cannot be measured before it operates.”

Unseen practice: A school plan estimates that a new reading programme could save ten minutes per lesson. Write an accurate status sentence before launch.

Explained answer 6

“The plan estimates a potential ten-minute saving per lesson.” It is not yet correct to write “the programme saves ten minutes” unless the school has measured the actual result.

Changed condition: A pilot later observes an average seven-minute saving in two classes. Report the observed pilot result separately from the original projection; the two numbers answer different questions.

What this teaches: Estimated, projected, planned and measured are not interchangeable rhetorical flourishes.

7. A small count converted into a precise percentage without context

Two pupils out of three in a fictional corridor interview say they enjoy longer lunchtime reading. A student reports “66.7 per cent of schoolchildren endorse the policy.” The denominator and recruitment method disappear.

Diagnose the claim leap: The percentage is arithmetically close to two thirds but implies a measured school-wide pattern that the three-person interview does not establish. False precision can strengthen the illusion of authority without improving information.

Supported revision: “Two of the three pupils interviewed in the corridor said they liked longer lunchtime reading. Their views cannot stand for the whole school without a suitable wider sample.”

Unseen practice: Five of six volunteers favour a club proposal. Write an accurate sentence and explain why it cannot be treated as the view of every member.

Explained answer 7

“Five of six volunteers who responded favoured the proposal.” Volunteers may not represent non-respondents, and the sample describes only six people. The writer needs no decimal places to make the finding useful.

Changed condition: Now a documented full membership survey becomes available. Compare its coverage and questions before deciding whether the original small interview can be used illustratively.

What this teaches: Exact arithmetic does not rescue an unearned population claim.

8. A survey answer treated as actual behaviour

A fictional form asks pupils whether they intend to borrow more books next month. Forty respondents choose yes. An essay concludes that forty pupils borrowed more books.

Diagnose the claim leap: Intended future behaviour and recorded past behaviour are different variables. The survey can tell us what participants said they planned, not whether those plans were later carried out.

Supported revision: “Forty respondents said they intended to borrow more books the following month. The survey does not establish what they subsequently borrowed.”

Unseen practice: A questionnaire asks whether parents would use an online booking option. Seventy select yes. Has the school demonstrated seventy uses? Repair the statement.

Explained answer 8

No. It has seventy affirmative stated intentions in that questionnaire. Actual usage would require a later booking record, with appropriate privacy and measurement safeguards.

Changed condition: When an anonymised booking total arrives, the writer may compare planned and observed uptake, while acknowledging the records may not identify the same individuals.

What this teaches: Words such as intend, expect, report, did and will mark genuinely different evidence.

Source casebook B — Counts, methods and numerical meaning

9. Average reported as every student’s experience

In an invented class trial, the average completion time falls from 18 to 15 minutes. The argument says every pupil completed the task three minutes faster. The report provides no individual times.

Diagnose the claim leap: A change in a group average does not determine the change for each participant. Some could improve greatly, some remain unchanged and some take longer, while the average still falls.

Supported revision: “The class’s average completion time was three minutes lower in the later attempt. The available summary does not show how each pupil’s time changed.”

Unseen practice: A library’s mean borrowing count rises from four to five books per member. Can we say every member borrowed one more?

Explained answer 9

No. The mean rose by one, but members’ individual changes may differ. To discuss whether most participants increased their borrowing, we need the underlying distribution or an appropriate additional summary.

Changed condition: A new graph reveals that half improved while others were unchanged or slower. The overall average and the distribution must both be represented without contradiction.

What this teaches: Aggregate language should not silently become individual-level language.

10. A median mistaken for the maximum

A school poster says the median reading time in a sample was thirty minutes. The student writes that no sampled pupil read longer than thirty minutes, using the median as a boundary.

Diagnose the claim leap: The median describes a central position in ordered data; it is not the longest observation. A vocabulary misreading has produced a false constraint.

Supported revision: “The sample’s median reading time was thirty minutes.” Without the underlying values, the note cannot identify the maximum, range or number of readers above a particular threshold beyond the median’s positional implication.

Unseen practice: A dataset has ordered times 10, 20, 30, 40, 80 minutes. Identify median and maximum and explain which word answers each question.

Explained answer 10

The median is 30 minutes, the central value. The maximum is 80 minutes, the highest recorded value. Writing “no one read more than 30 minutes” would contradict the supplied data.

Changed condition: Replace 80 with 200. The median remains 30 while the maximum changes sharply; this helps show why the two words should not be interchanged.

What this teaches: Statistical vocabulary carries a mathematical definition, not a persuasive mood.

11. Correlation called proof of causation

A fictional survey finds that pupils who practise instruments also report reading more. A student argues that instrumental practice causes reading habits. No design rules out family support, time availability or personal interests.

Diagnose the claim leap: The relation may be real in the described data, but the study has not isolated a causal mechanism. The writer should not simply replace “causes” with “contributes to” unless evidence supports even the weaker causal assertion.

Supported revision: “In this survey, instrumental practice and self-reported reading were associated. The design does not by itself establish that one caused the other.”

Unseen practice: Classrooms with more windows also report higher test scores. Write one association sentence and one alternative factor the study may need to consider.

Explained answer 11

“In the observed classrooms, window count was associated with higher scores.” Room size, building age or class allocation might also matter. The task does not claim any of those factors actually caused the pattern.

Changed condition: If a later credible experiment isolates a relevant effect, a causal claim may become possible within its population and conditions, still without implying a universal result.

What this teaches: A correlation can motivate a question without resolving it.

12. A first-hand account presented as a census

A single commuter writes a detailed diary about long queues at one bus stop over three days. An editorial uses the diary to claim that commuters across the entire country face the same daily delay.

Diagnose the claim leap: First-hand specificity is valuable evidence about the recorded experience. Rich detail does not automatically make it representative of every commuter, place or day.

Supported revision: “In a three-day diary, one commuter recorded long queues at this stop. The account provides a concrete example, but broader transport data would be needed to describe typical national conditions.”

Unseen practice: One student writes that the canteen queue took twenty minutes on Monday. Is “the queue always takes twenty minutes” supported?

Explained answer 12

No. The account supports the student’s reported Monday experience, assuming accuracy. A frequency claim would require repeated comparable observations, ideally with clear definitions of queue start and finish.

Changed condition: Additional independently collected records across weeks may show a pattern at that canteen. Scope the conclusion to the period and place actually measured.

What this teaches: An illustrative case can bring an argument to life without carrying the population inference.

13. Document date mistaken for event date

An archived article published in 2019 recounts a local exhibition that happened in 2012. The writer lists the exhibition as a 2019 event because the visible website date was taken as the event date.

Diagnose the claim leap: Sources carry publication, update, access and event dates. Those are distinct metadata fields. The wrong date can distort a historical comparison even when the article’s narrative is otherwise read correctly.

Supported revision: “The exhibition took place in 2012, according to an article published in 2019.” A complete citation can keep the two dates distinct where relevant.

Unseen practice: A museum page updated in 2026 describes a 1995 donation. What year belongs to the donation claim?

Explained answer 13

1995 is the described donation year. The 2026 update date tells readers when the page was revised, not when the donation occurred. Verify the original record if the exact year is consequential.

Changed condition: If a later correction changes the donation year, use the corrected information and preserve a note about which version was consulted.

What this teaches: A newer webpage is not necessarily evidence of a newer historical event.

14. A translated summary reported as verbatim English

A guest speaks in Malay during a fictional class interview. The editor translates the sense into English and places the resulting sentence in quotation marks without noting that it is a translation.

Diagnose the claim leap: Translation involves choices and may compress or clarify the original. The English wording is not the precise original speech. Readers deserve to know whether they are seeing a quotation in translation or a paraphrase.

Supported revision: “In an English translation, the guest said that the programme made the neighbourhood feel more welcoming.” Exact quotation marks require faithful translation and transparent identification as translated speech.

Unseen practice: A summary of a speech in another language says that a speaker was “furious”. The actual expression could mean upset or concerned. What limitation belongs beside the paraphrase?

Explained answer 14

The strength of the emotional term needs checking against the original phrase and context. The editor should not attribute the stronger emotion without a reliable basis; a cautious paraphrase may be appropriate.

Changed condition: A bilingual reviewer supplies a checked translation. Update the attribution and exactness claim, but do not pretend the English sentence was originally spoken in English.

What this teaches: Provenance includes language, editorial handling and speaker ownership.

15. A chart scale read without the baseline

A student looks at a bar chart whose vertical axis begins at 90 rather than zero. Small differences in recorded values appear dramatically large, and the essay says one group was “several times better”.

Diagnose the claim leap: The bars’ visual heights do not represent a ratio from zero. The numerical values, units and axis baseline determine the defensible comparison. A chart can communicate accurate numbers yet invite a misleading visual impression.

Supported revision: “The recorded values differed by four points on a scale whose axis begins at ninety. The graph’s appearance exaggerates the visual height difference relative to the numerical gap.”

Unseen practice: Two scores are 96 and 100. A truncated chart shows one bar twice as tall. Can the student say the second score is twice the first?

Explained answer 15

No. 100 is only four points higher than 96, not twice as large. The axis design accounts for the visual difference. The reader should calculate the actual comparison before writing.

Changed condition: If the chart is redrawn with a zero baseline, the figures are unchanged. Revise claims about the visual presentation while retaining the true numeric difference.

What this teaches: Graph vocabulary includes axis, unit, scale, baseline and denominator, not merely taller and shorter.

16. A missing figure quietly replaced by zero

A report table uses a dash where one school’s attendance figure was not submitted. A student treats the dash as zero and concludes that the school had no attendees.

Diagnose the claim leap: Missing and zero answer different questions. Zero is a measured count or valid recorded value; a missing mark may indicate no report. Substituting one for the other can create a dramatic but false comparison.

Supported revision: “The attendance figure for that school was not supplied in the table; it cannot be treated as zero without a stated data-coding rule.”

Unseen practice: A column displays 12, 14 and “not recorded”. May a pupil average the three entries by using zero for the missing one?

Explained answer 16

Not without a justified method and clear treatment of missing data. If the task seeks the average of recorded values only, the mean of 12 and 14 is 13, with a note that one entry was unavailable.

Changed condition: If documentation later confirms that “not recorded” was a coding error for an actual measured zero, update the calculation and document the correction.

What this teaches: Not available, not applicable, withheld and zero must not be collapsed into one label.

Source casebook C — How research design limits a conclusion

17. The survey question that nudges the answer

A fictional student poll asks, “Don’t you agree that our excellent new lockers should stay?” The essay reports that eighty per cent of respondents approved the lockers, without showing the leading question.

Why the claim exceeds the source: The percentage may correctly summarise the answers to that question, but the wording creates a risk that the question itself encouraged agreement. The writer should not present the result as an unqualified measurement of underlying preference.

Repair: “In a poll using a positively framed question about the lockers, eighty per cent of respondents selected agreement. The wording may have influenced how participants responded.”

New prompt: Compare “How useful are the new lockers?” and “Surely everyone finds our wonderful lockers useful?” Which question is more neutral, and what limitation belongs beside results from the second?

Answer with reasoning — case 17

The first is more neutral if appropriate balanced response choices are provided. The second encourages a favourable answer. Its results should be described as responses to a leading question, not as uncontaminated preferences.

What changes under new evidence: A later poll asks neutral questions with balanced options. Comparing the two is possible only if differences in sample, time and measurement are also considered.

Boundary retained: Survey vocabulary must preserve the question’s wording and the population invited to answer.

18. The self-selected respondents treated as random

An online form about school transport is posted in a club’s group chat. Pupils who care strongly about the issue are more likely to open it. The report refers to participants as a random sample of students.

Why the claim exceeds the source: A link open to volunteers is not the same as a probability sample selected with an appropriate random method. Describing it as random invents a methodological safeguard that the research did not use.

Repair: “Respondents chose to complete an online form shared in one club’s group chat. Their answers describe this volunteer group; they may not represent pupils who did not see or complete the form.”

New prompt: A student surveys whoever walks past their classroom at lunch. Is this necessarily a random sample of the entire year group?

Answer with reasoning — case 18

No. It is a convenience sample with location and time boundaries. Other pupils may be in lessons, canteen queues or off-site. A careful report names the selection method.

What changes under new evidence: If the school later uses a documented random selection from an appropriate roster, the new method may permit stronger population inference, subject to response and measurement limits.

Boundary retained: Random is a technical statement about selection, not a synonym for informal.

19. Non-response treated as disagreement

A team sends fifty invitations to comment on a proposed class project. Twenty reply; twelve favour it, eight oppose it. The writer says thirty-eight pupils oppose the proposal because they did not send a positive response.

Why the claim exceeds the source: Thirty pupils provided no answer. Their silence could reflect disinterest, absence, technical difficulty or many other reasons. It is not an observed negative vote.

Repair: “Of the twenty pupils who responded, twelve favoured and eight opposed the proposal. Thirty invitees did not respond, so their preferences are unknown.”

New prompt: A school notice receives three objections and no other messages from forty readers. Can a pupil conclude that thirty-seven people support it?

Answer with reasoning — case 19

No. Lack of an objection is not affirmative support. At most, three objections were recorded and the views of the others were not established through that channel.

What changes under new evidence: If a later confidential ballot obtains responses from all forty, the result can provide more direct evidence while remaining subject to how the choice was framed.

Boundary retained: Responded, abstained, not reached and unreported are separate states.

20. A comparison without the same question

Group A is asked whether lunch options are affordable; Group B is asked whether they are enjoyable. An essay compares the favourable percentages as though both groups answered the same satisfaction question.

Why the claim exceeds the source: The instrument measures different constructs. A higher percentage may reflect wording rather than a real group difference. The correction begins by preserving what each group was asked.

Repair: “Group A reported views on affordability, while Group B answered a question about enjoyment. Their response percentages cannot be compared as if they measured one identical judgement.”

New prompt: One class rates the difficulty of homework, another rates how interesting it is. Can we say the second class found homework easier?

Answer with reasoning — case 20

No. Interest and difficulty are distinct. The source has not directly measured relative difficulty in both classes. The writer can describe each result separately.

What changes under new evidence: A standardised follow-up asks both groups the same clearly defined question. Now comparison is more defensible, with relevant sample differences still noted.

Boundary retained: The words used in a survey are part of its measurement, not a decorative introduction.

21. A category defined differently in two years

An invented school annual report labels a pupil “active reader” in 2025 if they borrow at least one book, but in 2026 requires four books. The graph shows fewer active readers in 2026. The essay blames a decline in interest.

Why the claim exceeds the source: The classification threshold changed. The same behaviour can produce a different label across years, so the count cannot be interpreted as a straightforward trend in engagement.

Repair: “The reported number of ‘active readers’ fell under a stricter 2026 definition. The changed threshold prevents a direct comparison with the 2025 figure without recalculating under a common rule.”

New prompt: Last year’s club register counted attendance once per month; this year it counts attendance once per week. May raw participation percentages be compared?

Answer with reasoning — case 21

Not without making the periods and conditions comparable. Even identical words such as “attended” can hide different definitions and recording windows.

What changes under new evidence: If both years’ raw logs become available, a common measure may be computed with clear assumptions. That is a new analysis, not a simple correction of a spelling error.

Boundary retained: Defined terms must remain stable across the comparison a sentence makes.

22. A ranking mistaken for a measured improvement

A student says a school moved from third to second in an invented competition and therefore improved its performance by one third. The ranking changed, but the scores and other teams’ results are not given.

Why the claim exceeds the source: An ordinal position expresses order, not the proportional size of a difference. Another team may have declined while the school’s score stayed the same. A rank change alone cannot calculate magnitude.

Repair: “The school moved from third to second place in the competition. Without the underlying scores and criteria, the change in relative standing does not establish how much its performance improved.”

New prompt: A runner finishes fifth one year and fourth the next. Can we conclude that their speed improved by twenty per cent?

Answer with reasoning — case 22

No. Position is not a direct speed measure, and the participant field may change. Race times and comparable conditions are needed to discuss speed changes.

What changes under new evidence: If reliable times show a faster performance under similar conditions, the writer can discuss both the time difference and the ranking, separately.

Boundary retained: Rank, score, proportion and rate are different numerical languages.

23. A percentage increase read as percentage points

An invented attendance rate rises from forty to fifty per cent. The student calls it a ten per cent increase. That phrase can mean a relative change, which would be twenty-five per cent of the original forty, not ten.

Why the claim exceeds the source: The difference is ten percentage points. The relative percentage increase is ten divided by forty, or twenty-five per cent. The source-limitation check must protect the baseline from being silently swapped.

Repair: “Attendance rose by ten percentage points, from forty to fifty per cent, equivalent to a twenty-five per cent relative increase from the initial rate.”

New prompt: The share of pupils choosing a club rises from 20 per cent to 30 per cent. State the difference in percentage points and the relative percentage increase.

Answer with reasoning — case 23

The increase is ten percentage points. Relative to the original twenty per cent, the change is ten divided by twenty, or fifty per cent. Both are valid if correctly named.

What changes under new evidence: If the baseline changes to twenty-five per cent, the same ending rate of thirty represents five percentage points and a relative twenty per cent rise.

Boundary retained: Mathematical vocabulary prevents an impressive-sounding but incorrect comparison.

24. A mean concealed by an extreme case

Ten pupils complete a task in roughly ten minutes, while one needs a much longer period for an accessibility-related reason. A writer uses the mean alone to describe how long a “typical pupil” takes, then recommends forcing everyone to match it.

Why the claim exceeds the source: An average may be influenced by an outlying value and cannot represent every individual’s needs. Moreover, a distribution summary is not a reasonable basis for disregarding approved accommodations.

Repair: “The mean completion time describes this sample’s overall average; it does not establish a suitable deadline for every pupil. Individual variation and task conditions matter.”

New prompt: A group of five reading times is 8, 9, 10, 11 and 42 minutes. Compare mean and median and explain which better shows the central cluster.

Answer with reasoning — case 24

The mean is 16 minutes; the median is 10. The unusually long time affects the mean. A report should show the chosen statistic and not use either value to dictate a universal speed.

What changes under new evidence: If the longest time reflected a technical interruption rather than reading, check how the dataset defines completion before deciding whether to exclude or separately describe it.

Boundary retained: A useful limitation names the distribution and fairness consequences, not just the average.

Source casebook D — Corroboration, provenance and missing comparisons

25. A claim that uses “significant” without meaning

A class survey score rises from 72 to 73. The student calls it a “significant improvement” without explaining whether the term means practically important, statistically tested or simply noticeable.

Why the claim exceeds the source: Significant has everyday and statistical senses. A one-point difference might matter in context, but the writer must not imply a statistical result that was never calculated.

Repair: “The score increased by one point in this sample. The practical importance and uncertainty of that change have not been established by the available summary.”

New prompt: A reading rate changes from 24 to 25 items per minute. Rewrite “a highly significant gain” when no statistical test or repeated measurement is supplied.

Answer with reasoning — case 25

“The recorded rate rose by one item per minute in this observation.” A conclusion about reliable learning improvement would need appropriate repeat measures and analysis.

What changes under new evidence: If a carefully designed evaluation later supplies an uncertainty interval or statistical test, report what it actually measures while also considering real-world relevance.

Boundary retained: An impressive technical term should never be used merely as a stronger adjective.

26. A qualitative theme flattened into a headcount

A fictional researcher notes that several interviewees mentioned a confusing sign. A student writes that 70 per cent of residents thought the sign was confusing, although no count or denominator is supplied.

Why the claim exceeds the source: “Several” in a thematic summary is not a measured percentage. Interviews may be designed to explore experiences rather than estimate a population proportion.

Repair: “Confusing signage emerged as a concern in several interviews. The available thematic summary does not provide a percentage of residents who share that view.”

New prompt: A report says “many participants appreciated the quieter space” but includes no sample size. Can the writer conclude that a majority did?

Answer with reasoning — case 26

No. “Many” does not establish more than half of a defined sample. Retain the report’s cautious language or inspect original coded counts if available and appropriate.

What changes under new evidence: A later appendix lists exact codes and interview totals; the writer can describe proportions among participants while preserving qualitative sampling limitations.

Boundary retained: Themes, frequency counts and population estimates serve different research purposes.

27. A quoted expert assigned authority outside their field

A fictional designer explains why the labels on a school map are hard to read. An essay cites the same designer as proof that the new map has improved student mental health.

Why the claim exceeds the source: Expertise is task-specific. The source may be directly relevant to typography and wayfinding while lacking evidence for psychological outcomes. Credentials cannot substitute for the right outcome measures.

Repair: “The designer argues that the map’s lettering could be easier to read. The interview does not establish whether the redesign changed pupils’ wellbeing.”

New prompt: A mathematician explains a chart’s calculation. May that opinion alone prove the policy depicted is socially fair?

Answer with reasoning — case 27

No. The mathematical explanation can help assess the chart, but fairness requires appropriate criteria, affected perspectives and evidence about consequences.

What changes under new evidence: If a separate evaluation examines accessibility and wellbeing outcomes, the writer can combine findings while keeping each source attached to the claim it actually supports.

Boundary retained: Source credibility depends on relevance as much as reputation.

28. An anonymous source with a hidden identity claim

A school publication uses a fictional anonymous interview about a club dispute. The editor describes the speaker as “a teacher with direct access to the decision”, even though the notes only establish that an unidentified person supplied an account.

Why the claim exceeds the source: Anonymity protects a name, but cannot justify inventing a role, access level or privileged knowledge. In real schoolwork, safeguarding and consent procedures govern whether such material should be used at all.

Repair: “An anonymous source gave this account. The available record does not establish the source’s role or first-hand access, so the claim needs independent verification before publication.”

New prompt: A message from an unnamed account alleges that a timetable is changing. Write a responsible report status without repeating the allegation as established fact.

Answer with reasoning — case 28

“An unverified message has circulated about a timetable change. The school has not confirmed the change through its official channel.” Publication decisions should follow school rules, especially when real people are involved.

What changes under new evidence: An authorised notice later confirms the new schedule. Report the confirmation from that notice rather than upgrading the anonymous message retroactively into verified authority.

Boundary retained: Anonymous is not a synonym for knowledgeable or unreliable; the evidential role must be established.

29. Two dependent articles mistaken for two independent sources

A student cites an education blog and a news summary as independent evidence for a claim. Both reproduce the same short press release without doing any additional reporting.

Why the claim exceeds the source: Two links do not necessarily provide independent corroboration. If one is copied from the other or both originate in the same release, their agreement may only repeat the original assertion.

Repair: “Both articles repeat the same underlying announcement. They should not be counted as two independent checks of its factual accuracy.”

New prompt: Three websites publish identical statistics credited to one association. Does agreement among the sites establish three separate datasets?

Answer with reasoning — case 29

No. Unless they independently collected or verified data, the sites may all be relying on one association’s figures. Trace the provenance and evaluate that original source.

What changes under new evidence: An independently collected school record later confirms part of the claim; note exactly which part gains corroboration and what remains untested.

Boundary retained: Corroboration concerns independence and method, not just the number of URLs.

30. The edited source whose old version disappears

A web page once listed registration closing on Friday. An editor updates it to Thursday without a visible change log. A student’s earlier notes quote Friday; the current page shows Thursday.

Why the claim exceeds the source: The apparent contradiction may reflect document version, not careless reading. The writer must separate what was recorded at the time from what the present source says, without pretending to know why it changed.

Repair: “In the version consulted earlier, the page listed Friday; the current version lists Thursday. The change’s reason is not documented in the material available to us.”

New prompt: An event leaflet photographed last week says Room A, but today’s official website says Room C. What should a practical school update tell readers?

Answer with reasoning — case 30

Use the latest verified authorised information for the current action, while explicitly noting that the old location has been superseded if confusion is likely. Do not guess whether the first leaflet was an error or the venue simply changed.

What changes under new evidence: When an official correction note gives a reason and timestamp, it can be added with attribution; preserve the distinction between recorded versions.

Boundary retained: Version control and source provenance prevent yesterday’s true quotation from masquerading as today’s instruction.

31. The absent comparison group that weakens causal language

A study of one class introduces a new study technique and records improvement. A pupil writes that the method is more effective than ordinary teaching, but no comparable class or alternative was evaluated.

Why the claim exceeds the source: Improvement over time can occur for many reasons. To compare methods, the design needs information about an appropriate alternative under relevant conditions. The source may demonstrate observed improvement, not superiority.

Repair: “The class’s recorded performance improved during the period when the technique was used. The available report does not establish that it outperformed another approach.”

New prompt: A school introduces extra reading time and later sees higher scores. Write a cautious sentence that does not assert other support was absent.

Answer with reasoning — case 31

“Scores rose after the extra reading period began. Because other influences were not isolated here, the figures alone do not show how much the reading period contributed.”

What changes under new evidence: If a well-described comparison later becomes available, revise the evaluation within the populations and conditions actually tested.

Boundary retained: A source-limitation note asks which comparison would justify the particular causal or superiority claim.

32. Uncertainty expressed as an invented exact probability

A student reads that a school project “may be completed in November” and writes there is a 70 per cent chance it will finish on time. The report supplies no formal probability forecast or model.

Why the claim exceeds the source: Quantifying confidence can look professional while fabricating information. “May” expresses possibility but does not necessarily encode a specific percentage. The writer has introduced an unsupported precision.

Repair: “The report identifies November as a possible completion date, but it does not quantify the likelihood of meeting it.”

New prompt: A teacher says it is “likely” that a class visit will go ahead. May the pupil translate this into an 85 per cent probability?

Answer with reasoning — case 32

No, not from that wording alone. The teacher may have an informal judgement; a precise probability needs appropriate definition and basis. A faithful report says that the visit is regarded as likely, with the source named.

What changes under new evidence: An organiser later publishes a model-based numerical estimate. Report the estimate as theirs, explain its assumptions if known and do not treat it as a guaranteed event.

Boundary retained: Honest uncertainty is more useful than decimal precision without evidence.

Source casebook E — Reading news, images and persuasive framing

33. A headline that silently changes the verb

An invented news headline reads “Students Reject Hall Plan”. Inside the article, twelve pupils question whether the hall can handle a larger audience but have not voted to reject anything. The headline makes their caution sound like a settled decision.

Where the interpretation goes too far: Objecting to one condition and rejecting a whole plan are different actions. A student checking sources must look for the decision verb in the original evidence, not borrow a stronger one from the headline.

Safe, useful rewriting: “Twelve pupils raised concerns about the hall’s capacity. The article does not establish that they formally rejected the proposed event.”

Practice: A school newsletter says “Families Demand Longer Breaks” but quotes two parents asking whether a trial is possible. Replace the headline with a defensible one.

Reasoned response for case 33

“Two Parents Ask About Trial of Longer Breaks” matches the quoted request. “Families demand” enlarges the number and intensifies the force beyond what the article supplies.

New evidence: If a formal petition later gathers signatories explicitly requesting the change, the report can describe that action and count. The earlier two-parent enquiry remains a different event.

Meaning to retain: Verbs such as asked, proposed, questioned, opposed, rejected and decided must match documented action.

34. An opinion column cited as a factual announcement

A school blog column argues that students would benefit from a later start. Its author discusses personal experience and recommendations. A pupil treats it as proof that a later start has been officially approved.

Where the interpretation goes too far: An editorial may give relevant reasoning, but its genre and authorial position are not the same as an official decision notice. The student confuses advocacy with implementation.

Safe, useful rewriting: “The columnist supports a later start and offers reasons for it. The article does not establish that the school has adopted the proposal.”

Practice: A student writes “The school will ban plastic bottles” after reading an opinion piece urging a ban. What must be checked?

Reasoned response for case 34

Look for an authorised policy statement or notice. Until then, report that the author recommends a ban. A recommendation is a possible future action, not an enacted requirement.

New evidence: An official statement later confirms a phased policy. The writer can now describe the policy, dates and exceptions from that statement, without treating the earlier column as the decision authority.

Meaning to retain: Genre identification protects readers against treating a position as an institution’s action.

35. A fact-check score mistaken for universal truth

A classroom checklist gives an article four ticks out of five for citation, date, author and clarity. The student’s report says the article is “80 per cent true”. The ticks assessed presentation criteria, not all factual claims.

Where the interpretation goes too far: A rubric’s score describes performance on named criteria. Unless designed and validated for factual accuracy, it cannot be converted into a probability of truth or a percentage of claims that are correct.

Safe, useful rewriting: “The article met four of five checklist criteria. This assessment does not mean that eighty per cent of its factual statements have been verified.”

Practice: A source receives “good” for readability but “uncertain” for evidence. May an essay call it highly reliable overall solely because it is easy to read?

Reasoned response for case 35

No. Readability and accuracy are distinct. The student should explain which criterion was met and which remains unresolved, especially for consequential factual claims.

New evidence: If a separate claim-by-claim audit checks all evidence and records a method, the two assessments can be considered together without collapsing them into one misleading percentage.

Meaning to retain: Scoring language must name the construct that was actually evaluated.

36. A screenshot without the previous message

A student shares a screenshot of a fictional group chat saying, “That plan is ridiculous.” Their commentary says one classmate insulted another’s whole project. The visible screen does not include the earlier message or the reply after it.

Where the interpretation goes too far: The words are visible but the object of criticism and conversational context are missing. The writer should not reconstruct motive or target merely because the sentence sounds harsh.

Safe, useful rewriting: “The screenshot shows a participant writing, ‘That plan is ridiculous.’ Without the surrounding exchange, the record does not show precisely which plan was being discussed or how others responded.”

Practice: A clipped email says “Please cancel it” but not what “it” refers to. Write a limitation note.

Reasoned response for case 36

The message contains a cancellation request with an unresolved object. The missing previous correspondence is needed before a writer can identify the event, meeting or order involved.

New evidence: The full thread later identifies a room booking as the object. Revise the description to name the booking while preserving the original quote’s wording.

Meaning to retain: Reference words depend on context; screenshots can hide their owners.

37. A literary narrator confused with the author

In a short fictional story, a narrator says, “Everyone in the village despised the newcomer.” The student writes that the author believed every villager was prejudiced. The narrative voice is one constructed perspective inside the work.

Where the interpretation goes too far: The narrator’s assertion may be reliable, mistaken, ironic or limited. Treating it as the author’s personal opinion crosses a literary-source boundary. Close reading must identify who speaks.

Safe, useful rewriting: “The narrator presents the village as hostile to the newcomer. The wording may reveal the narrator’s generalising perspective; it does not by itself establish the author’s personal belief.”

Practice: A character calls the city “cold and cruel”. Does this prove that the novelist personally dislikes cities?

Reasoned response for case 37

No. It reveals a character’s or narrator’s viewpoint within the text, depending on how it is presented. The writer can analyse language and perspective without attributing a belief to the real author.

New evidence: Other passages later undermine the narrator’s judgement. The student’s interpretation should evolve to discuss narrative unreliability or irony if the text supports that reading.

Meaning to retain: Attribution matters inside fiction as much as it does in news reporting.

38. A diary entry made into a complete historical record

A fictional traveller’s diary says that the town market was closed on the one morning they visited. The student concludes the market never opened on Mondays in that historical period.

Where the interpretation goes too far: One personal observation establishes at most the traveller’s described experience, and even the date may need checking. A broad schedule claim requires additional contemporary evidence.

Safe, useful rewriting: “The traveller recorded finding the market closed during that visit. The diary does not establish the market’s normal weekly opening pattern.”

Practice: A letter from one family mentions bread shortages during a particular winter. What can it support about all families in the region?

Reasoned response for case 38

It offers an account of that family’s reported circumstances. A claim about regional prevalence would need additional records, context and attention to whose voices were preserved.

New evidence: An archived market rule and several dated shop ledgers become available. The student can triangulate a more precise pattern, while explaining where sources agree or conflict.

Meaning to retain: Primary sources are not omniscient; they offer particular access to a particular time.

39. A public policy stated without the exception

An invented rule says, “All visitors must register at the entrance, except emergency responders carrying out official duties.” A student’s summary says every person entering the building must register without exception.

Where the interpretation goes too far: The writer has changed the policy’s scope by dropping a material condition. The word “all” appears in the rule but does not cancel the explicit exception following it.

Safe, useful rewriting: “Visitors are generally required to register at the entrance; the rule states an exception for emergency responders on duty.”

Practice: A classroom rule says phones must be kept away during tests unless an approved accessibility arrangement requires one. Write a precise summary.

Reasoned response for case 39

“Phones are normally kept away during tests, with the stated exception for approved accessibility arrangements.” This preserves a necessary distinction without inventing which pupils qualify.

New evidence: If the policy changes and removes or adds an exception, check the authorised latest text and update the summary accordingly.

Meaning to retain: Exceptions belong to the meaning of a rule, not its optional fine print.

40. A recommendation confused with an obligation

A fictional guide says students “should consider bringing a reusable water bottle” on a trip. A pupil writes that everyone “must” bring one and will be penalised if they do not.

Where the interpretation goes too far: Advice, permission, obligation and prohibition have different strengths. The invented penalty is unsupported. A strong sentence should preserve the modal and its authority.

Safe, useful rewriting: “The guide recommends that pupils consider bringing a reusable bottle. It does not state that this is compulsory or that noncompliance will be penalised.”

Practice: A notice says “you may use a pencil for rough notes”. Does that oblige everyone to use one? Repair the wording.

Reasoned response for case 40

No. It permits an option. A summary might read, “Pupils are allowed to use a pencil for rough notes,” without presenting it as mandatory.

New evidence: A later assessment instruction says “You must use black ink for final answers.” Explain how that separate requirement changes the allowed actions for final responses.

Meaning to retain: Modal verbs carry practical authority boundaries even in ordinary school notices.

Source casebook F — Text, argument and public interpretation

41. A scientific diagram described as a photograph

An invented textbook shows a coloured schematic of water flowing through a filter. A pupil captions it, “This photograph proves the filter removes every harmful substance.” The drawing is an explanatory representation, not a direct record of a test.

Where the interpretation goes too far: Illustrations can teach mechanisms but their status matters. A schematic may simplify scale, omit side processes and show intended function without testing outcomes. A photograph would also not establish universal effectiveness by itself.

Safe, useful rewriting: “The diagram illustrates the proposed sequence of filtration. Its labels do not provide direct evidence that the system removes every contaminant.”

Practice: A science poster uses arrows to show energy transfer in a model. Can the arrows prove the real device has no energy losses?

Reasoned response for case 41

No. The arrows represent the intended conceptual relationship. Measurements and an appropriate model are needed for claims about actual efficiencies or losses.

New evidence: An experimental report later supplies removal rates for named substances under defined conditions. The student can report those measured outcomes, not claim all possible contaminants disappear.

Meaning to retain: Representation, simulation, observation and measurement are different kinds of evidence.

42. An unlabeled axis assigned a convenient unit

A graph displays numbers 20, 40 and 60 but omits the vertical axis label. A student describes them as percentages because this would support a point about success rates.

Where the interpretation goes too far: The visible numerals cannot by themselves establish whether the values represent people, minutes, scores or percentages. The missing unit prevents a faithful quantitative claim.

Safe, useful rewriting: “The chart displays values increasing from twenty to sixty, but the missing axis label prevents us from identifying the measurement or interpreting the practical size of the change.”

Practice: A table lists 150 beside one group and 160 beside another, without units. Can the writer say the second group is ten kilograms heavier?

Reasoned response for case 42

No. The figures differ by ten numerical units, but the measured quantity is unspecified. Request the legend, table notes or original dataset before naming kilograms.

New evidence: When a source note confirms the axis shows minutes, revise the description and any derived conclusion that assumed percentages.

Meaning to retain: A missing unit is not a blank space that rhetoric may fill.

43. A pie chart whose percentages do not add up

A fictional infographic labels four slices 35, 30, 25 and 20 per cent. These values total 110 per cent, yet an essay repeats the chart as if it were a coherent partition of one whole.

Where the interpretation goes too far: A basic arithmetic check reveals an inconsistency. The problem might stem from rounding, overlapping responses or a design error, but the graphic does not explain which. The student should not invent a correction.

Safe, useful rewriting: “The displayed percentages total 110 per cent. Without documentation explaining overlaps or rounding, the chart cannot be treated as a simple division of a single whole.”

Practice: A survey allows pupils to choose several activities. Percentages for reading, music and sport total 140. Must the chart be wrong?

Reasoned response for case 43

Not necessarily. If responses overlap, each percentage may refer to pupils selecting that activity, and totals can exceed 100. The caption should disclose the multi-select design and denominator.

New evidence: Documentation arrives confirming each respondent could select more than one activity. The student may keep the figures but must change the interpretation from exclusive slices to overlapping choices.

Meaning to retain: Numerical inconsistencies require checking the data structure, not just the graphic style.

44. A selected quote made to stand for the whole interview

A transcript shows a student describing both benefits and costs of an after-school programme. The published poster extracts only “I made new friends” and calls the interview overwhelmingly positive.

Where the interpretation goes too far: The selected quotation may be accurate in isolation while misrepresenting the overall balance of the source. Evaluation should consider whether omitted context changes the reasonable interpretation.

Safe, useful rewriting: “The poster highlights one positive statement from the interview. The full transcript also contains reservations, so the quotation alone does not establish an unqualified endorsement.”

Practice: An interviewee calls a class trip enjoyable but exhausting. A student uses only “enjoyable” as evidence that there were no difficulties. What is missing?

Reasoned response for case 44

The exhaustion remark qualifies the assessment. A faithful summary might say the trip was enjoyed but demanding, retaining both aspects instead of erasing the speaker’s limits.

New evidence: A later interview asks whether the person would join again and receives a clear yes. That new answer supports willingness to return, but does not erase the earlier fatigue.

Meaning to retain: Selection and omission can alter a source’s practical meaning even without changing a quoted word.

45. A list of reasons misrepresented as a ranking

A public notice lists three reasons for repairing a footpath: access, drainage and appearance. The student writes that appearance is the principal concern because it appears last in the list.

Where the interpretation goes too far: Order of presentation is not necessarily a priority ranking. Unless a source identifies “main”, “most important” or a weighting, assigning hierarchy adds a claim.

Safe, useful rewriting: “The notice lists access, drainage and appearance as considerations. It does not rank them by importance.”

Practice: A project brief mentions cost, ease of cleaning and durability. May a pupil say durability was the highest priority because it appeared first in their own notes?

Reasoned response for case 45

No. The brief’s explicit criteria or later evaluation would need to establish priority. Otherwise list the factors without imposing a ranking.

New evidence: A scoring table later assigns 50 per cent to durability. Now the student may state it was weighted most heavily in that assessment, provided the table is interpreted correctly.

Meaning to retain: Listed, prioritised, ranked and weighted are distinct descriptions.

46. A summary that merges contradictory accounts

One fictional attendee says a debate began at three; another remembers 3.20 pm. The student writes that both witnesses confirm a 3.10 pm start by averaging the recollections.

Where the interpretation goes too far: Conflicting memories are not automatically resolved by arithmetic. An average time was not observed by either witness. The writer has manufactured a consensus.

Safe, useful rewriting: “The two attendees gave different recollections of the starting time: 3 pm and 3.20 pm. The precise start remains unresolved by these accounts alone.”

Practice: Two historical diaries give different arrival dates. Can the writer choose the midpoint to create one date?

Reasoned response for case 46

No. Investigate why accounts differ, including dates, timekeeping, records and possible copying errors. Until resolved, attribute both versions or state the uncertainty.

New evidence: A dated video recording becomes available and supports a 3.12 pm start. The note may report that evidence and explain why the accounts were approximate without calling either witness dishonest.

Meaning to retain: Disagreement is information to examine, not a hole to fill with a conveniently averaged answer.

47. One word of an oral response detached from its explanation

During an invented spoken discussion, a pupil says a school initiative was “effective in raising awareness, but not yet effective in changing habits.” A peer writes that the speaker called it effective, without preserving the two different outcomes.

Where the interpretation goes too far: Effectiveness is always effectiveness at something. The source distinguished awareness from behaviour; removing the object makes the evaluation misleading.

Safe, useful rewriting: “The speaker thought the initiative had increased awareness but said there was not yet evidence of changed behaviour.”

Practice: A student says a poster was clear to read but failed to persuade volunteers to sign up. Summarise its strengths and limits in one sentence.

Reasoned response for case 47

“The poster was judged readable, but it did not persuade the group to register.” The summary names two different measures instead of declaring the poster simply good or bad.

New evidence: A later registration campaign shows greater uptake. The student should ask whether the poster changed or other factors were introduced before assigning cause.

Meaning to retain: Evaluative adjectives should carry a named outcome or criterion.

48. False balance between unequally supported claims

A student finds a well-documented school record and an unsupported anonymous rumour, then writes that the two accounts are “equally valid perspectives” simply to sound balanced.

Where the interpretation goes too far: Fairness means representing relevant views accurately; it does not require assigning equal evidential weight to every statement. A limitation note should say what is verified and what remains unsubstantiated.

Safe, useful rewriting: “The dated school record supports the stated schedule, while the anonymous claim lacks corroborating detail. Both can be reported as claims in circulation, but they do not carry the same evidential support.”

Practice: Two online posts conflict: one links a signed official notice, the other only says “my friend heard”. Should a report call them equally confirmed?

Reasoned response for case 48

No. The official notice directly documents the schedule. The friend’s account is hearsay unless independently checked. If the notice might be outdated, verify its current version rather than assigning fake equality.

New evidence: A newer official correction later supports the previously unverified timing. Update the current conclusion on that new evidence; being initially uncorroborated does not make a claim permanently false.

Meaning to retain: Evidence-weighted fairness is not the same as neutralising every difference.

Source casebook G — Calibrating English writing and spoken argument

49. Situational writing: a promise the prompt never authorised

A Secondary 4 pupil writes an invented school event proposal. The visual brief says the organisers are considering a change to seating. The student’s draft says, “We guarantee every attendee a front-row seat.” Neither capacity nor approval is established.

Claim diagnosis: Persuasive writing can propose benefits, but it must not invent a promise that an institution or organiser has not made. The problem is authority and scope, not merely the adjective “guarantee”.

Improved wording: “We propose arranging seats so that more visitors can see the stage clearly, subject to the hall’s capacity and safety requirements.” This clearly identifies a proposal and leaves operational approval to the responsible team.

Try it: A club notice asks for ideas to improve access to a workshop. Write one practical suggestion without claiming the school has already committed money or staff.

Answer and evidence boundary 49

“The club could provide a short written guide alongside the workshop demonstration, if the organiser agrees.” The suggestion is workable and conditional. “The school will employ extra coaches” would require authorisation.

If the record changes: The organiser later confirms an allocated budget and one additional helper. The writing can incorporate those verified commitments with their limits rather than converting them into unlimited support.

Transfer principle: The vocabulary of proposal, approval and commitment should follow the prompt’s actual permissions.

50. Argument writing: personal experience made into a rule

A writer says, “I worked better in silence last week, so schools should ban collaborative study.” The experience is relevant to the author’s preference but cannot establish the best arrangement for every learner and task.

Claim diagnosis: An argument may open with personal experience without treating it as a controlled comparison or population estimate. The student must distinguish a viewpoint from evidence of general effectiveness.

Improved wording: “My own work felt easier to organise in silence. This suggests that quiet study spaces matter for some tasks, although it does not show that group study should be prohibited.”

Try it: A student prefers handwritten notes and says digital notes never help learning. Repair the conclusion without denying the preference.

Answer and evidence boundary 50

“Handwriting helps me organise my notes, but this experience alone does not show that digital methods are ineffective for other students or tasks.” Further comparisons could address specific outcomes.

If the record changes: A classroom study compares methods for a defined task. Integrate its findings by specifying the task and population instead of claiming one universal method.

Transfer principle: Personal experience can be vivid support for a limited claim without becoming a universal law.

51. Comprehension: a source’s “some” becomes “all”

An informational passage says, “Some households in the pilot reported lower electricity use.” The pupil’s summary states every participating household saved energy, perhaps because the paragraph sounds positive overall.

Claim diagnosis: Summaries are not allowed to upgrade quantifiers. This is a direct reading error: the writer has made the source’s population more uniform than the original sentence permits.

Improved wording: “Some households in the pilot reported using less electricity.” If the passage supplies a count or time period, include it when relevant; otherwise do not invent one.

Try it: A story report says “Several visitors asked for signs in more languages.” Rewrite an answer that says “all visitors demanded translation”.

Answer and evidence boundary 51

“Several visitors requested more multilingual signage.” The revision restores the source’s limited group and softer speech act. It also avoids attributing a demand that the text does not show.

If the record changes: If a later survey obtains a majority preference, the report can state the sampled majority while preserving the difference from an absolute all.

Transfer principle: Compression should remove unnecessary detail, not the word that controls a claim’s scope.

52. Comprehension: negative evidence mistaken for positive proof

An invented passage says, “No complaints were logged during the trial.” The writer concludes that no participant was dissatisfied. A complaints register cannot directly record unreported feelings.

Claim diagnosis: Absence of a recorded complaint may reflect many processes, including whether participants knew how to complain or wanted to. The source may establish register status, not universal satisfaction.

Improved wording: “The trial’s complaints log contains no recorded complaints. This does not establish that every participant was satisfied.”

Try it: A school library received no written complaints about a new shelf layout. Can the pupil claim all readers found it easy?

Answer and evidence boundary 52

No. The absence of complaints is not a direct measure of accessibility or ease of use. Observations or a well-designed feedback channel might add evidence.

If the record changes: A follow-up usability test finds recurring search difficulties. The complaint log and the test can both be accurately described; one does not erase the other’s different measure.

Transfer principle: A source’s silence should not be promoted into universal affirmation.

53. Oral discussion: the strongest verb chosen for effect

A student discusses a fictional video about recycling and says “the campaign eradicated waste”. The video shows a small reduction in bins’ visible litter during one event.

Claim diagnosis: Eradicated means removed completely. A visible reduction at one time does not establish elimination across a site or after the event. The stronger word damages the precision of the point.

Improved wording: “The video suggests there was less visible litter in the filmed area after the activity.” The speaker can still praise the effort while asking whether the reduction lasted.

Try it: An advertisement says a community initiative “eliminates loneliness” but gives only two positive interviews. Give a respectful, evidence-proportionate response.

Answer and evidence boundary 53

“The interviews describe positive experiences for two participants, but they do not establish that the programme eliminates loneliness for everyone.” The critique targets the claim, not participants’ sincerity.

If the record changes: A long-term evaluation reports improvements on a specified wellbeing measure. The speaker may discuss those outcomes without converting a measured score into the claim that all loneliness disappears.

Transfer principle: Powerful verbs should be earned, not selected merely to sound decisive.

54. Oral disagreement: caution heard as hostility

A partner says a school app “must be safe because it is popular”. A pupil replies “There is no evidence for safety,” intending a careful challenge but sounding as though no evidence could ever exist.

Claim diagnosis: The premise “popular therefore safe” is unsupported, yet the blanket denial may go too far. A good oral question can challenge the inference without pretending to have reviewed all available evidence.

Improved wording: “Popularity alone does not establish safety. What information about privacy, access and testing is available?” The question identifies the missing connection.

Try it: A classmate says a product is reliable because many people have bought it. Form a fair response in one or two sentences.

Answer and evidence boundary 54

“High sales may show interest or reach, but they do not directly measure reliability. Have we seen testing results or consistent user reports?” This acknowledges one observation and asks for relevant evidence.

If the record changes: An independent reliability test later becomes available. The argument can now use its measured outcome within specified conditions rather than treating the sales number as proof.

Transfer principle: Respectful challenge language protects reasoning on both sides of a conversation.

55. A data table summarised without its denominator

An invented report compares ten damaged books in a library that lends 100 books with fifteen damaged books in a library that lends 1,000. The student declares the second library less careful because it has more damaged books in absolute count.

Claim diagnosis: The relevant rate may differ from the raw count. In this made-up example, ten in 100 is 10 per cent, while fifteen in 1,000 is 1.5 per cent; the second has more cases but a lower share.

Improved wording: “The second library recorded more damaged books in absolute number but a lower damage proportion relative to loans. A fair comparison should also check how damage is defined and recorded.”

Try it: A small club has four absent members out of twenty. A large club has ten absent out of a hundred. Which has the higher absence proportion?

Answer and evidence boundary 55

The small club has 20 per cent absent; the large has 10 per cent. The raw count is greater for the large club but its rate is lower. Neither statistic explains the reasons for absence.

If the record changes: If loan counts turn out to refer to transactions rather than distinct books, explain the denominator’s meaning before interpreting the rate.

Transfer principle: Counts and rates tell different stories; a limitation note should name the denominator.

56. A chart comparing unequal time periods

A report shows 120 books borrowed in one month and 200 borrowed across three months. The student claims borrowing increased from 120 to 200 after a reading campaign.

Claim diagnosis: The observation windows differ. The larger total covers three times as much time, so raw counts cannot establish an increase in rate. A comparable period or standardised measure is needed.

Improved wording: “The report lists 120 loans in one month and 200 over three months. Without aligning time periods, these totals do not establish that borrowing increased.”

Try it: A school collects 40 paper bags on Monday and 100 across Monday to Friday. Is 100 evidence of a faster daily collection rate?

Answer and evidence boundary 56

No. The five-day total covers more days. If collections were otherwise comparable, an average of twenty per day across the week could be compared with Monday’s forty; the calculation must retain the time unit.

If the record changes: When daily data becomes available, identify whether the campaign coincided with changes in daily rate, rather than only changes in accumulated total.

Transfer principle: Per day, per participant and over a period are essential vocabulary components.

Source casebook H — Transfer to Humanities, Science and real decisions

57. A before-and-after comparison that changed its participants

A school publishes a before-and-after graph for an optional club programme. Fifty students answered the first survey; thirty different students answered the second. The essay treats the averages as repeated measures on the same pupils.

Claim diagnosis: Two independent samples and a repeated-measures panel answer different questions. Without matching identifiers or a defined cohort, the writer cannot claim each participant improved.

Improved wording: “Average responses differed between two survey rounds with different participants. The figures do not show that the same pupils changed their views.”

Try it: Twenty pupils answer a class survey in January and twenty-five answer in June. May a pupil calculate every participant’s opinion change by subtracting the means?

Answer and evidence boundary 57

No. The means may be compared as samples with appropriate caveats, but individual change requires matched participants and consistent measurement.

If the record changes: A subset of matched anonymous codes later appears. Analyse the matched subset separately rather than treating all respondents as a single unchanged group.

Transfer principle: Cohort, respondent, sample and individual are distinct statistical owners.

58. A precise quote from a fictional character used as real-world advice

A novel’s wise mentor says “Never apologise before you win.” A student’s persuasive essay cites the line as evidence that apologising is always harmful in schools.

Claim diagnosis: Fictional dialogue can illustrate a character’s belief or thematic tension. It is not empirical evidence for a general behavioural claim, even when the wording sounds memorable.

Improved wording: “The mentor’s line presents a competitive view of apology in the novel. Its impact may depend on whether later events support or challenge that attitude.”

Try it: A character boasts that they never study and always succeed. Can the line prove that homework is unnecessary?

Answer and evidence boundary 58

No. It reveals the character’s claim or persona. The narrative might expose irony or unreliability. Research or appropriate real-world evidence would be required to argue about homework’s general effect.

If the record changes: If the plot later shows the mentor apologising meaningfully, the interpretation must reflect that narrative development rather than treating the line as the author’s settled doctrine.

Transfer principle: Literary evidence supports textual analysis, not automatically a universal social recommendation.

59. A first-person confession with missing context

An anonymous online comment says “I cheated once” without naming what action, occasion or rule is being discussed. The writer uses it as evidence that a particular group routinely cheats on exams.

Claim diagnosis: One unverified ambiguous statement cannot justify a group accusation, a frequency claim or identification of real people. This situation especially requires ethical and safeguarding caution.

Improved wording: “An anonymous commenter used the word ‘cheated’, but the available record does not establish the context, event or prevalence. The comment should not be used to accuse a named group.”

Try it: A message says “We broke the rule” without identifying the rule or the people included in “we”. Write a limitation note that protects both uncertainty and privacy.

Answer and evidence boundary 59

The message refers to an unspecified rule and group. The details are insufficient to identify an incident or attribute responsibility; seeking an appropriate authorised source may be necessary.

If the record changes: A later verified official account identifies a specific rule breach. The publication should rely on that evidence within its legitimate scope, not guess identities from the earlier ambiguous message.

Transfer principle: Ethical restraint is part of precision when real people might be harmed.

60. A proportion interpreted without its base year

A headline says exports rose 50 per cent, but the figure is compared with an exceptionally low previous year. The student concludes the current level is a record high. The stated growth rate does not establish the historical maximum.

Claim diagnosis: A large change from a low base may still produce a value below earlier years. A claim about records requires the full relevant series and consistent definitions.

Improved wording: “Exports rose 50 per cent compared with the specified previous year. The percentage increase alone does not show whether the current level exceeded earlier peaks.”

Try it: A school club doubles from five to ten members after reopening. May a pupil claim it has never been larger?

Answer and evidence boundary 60

No. Doubling is a relative change from five. An earlier club may have had thirty members. Historical records are needed for “record membership”.

If the record changes: If a complete verified series shows ten is indeed the maximum, the headline may become defensible with a clear record period.

Transfer principle: Relative growth, absolute level and historical record must be kept separate.

61. A local finding described as national policy

A student reads that one school has introduced a special timetable. Their oral presentation says all schools in Singapore now use it. The original report was a local school notice.

Claim diagnosis: Place and authority have changed. A school-level timetable cannot establish a nationwide MOE rule unless an authoritative source clearly states that policy.

Improved wording: “The named school introduced this timetable. The notice does not establish a requirement for every Singapore school.”

Try it: A local museum adds free entry for students. Can a class report say all national museums have made entry free?

Answer and evidence boundary 61

No. The finding concerns the museum in the source. A wider policy requires checking the relevant national or institutional authority and its eligibility conditions.

If the record changes: A later official national announcement may cover more institutions; describe its actual scope and exceptions without automatically assuming every venue is included.

Transfer principle: Local, institutional, regional and national are meaningful geographic and administrative limits.

62. A visual text with an invisible eligibility condition

A poster advertises “Free workshops for pupils,” while small print restricts registration to pupils from specified participating schools. A student’s announcement says any pupil can register.

Claim diagnosis: The main headline is real but incomplete. The eligibility note qualifies the offer. A reader who acts on the simplified message may waste time or face disappointment.

Improved wording: “The workshops are free for eligible pupils from the participating schools listed in the poster’s conditions.”

Try it: An event flyer says “Everyone welcome” but a footnote restricts the evening session to ticket holders. Write a practical summary for somebody planning to attend that session.

Answer and evidence boundary 62

“The evening session requires a ticket despite the general invitation on the flyer.” The writer should explain where ticket information is found when verified, rather than inventing a registration link.

If the record changes: If the organisers later open the session to all, the notice can be updated to reflect the change, with attention to the correct event date.

Transfer principle: Source evaluation includes captions, footnotes, definitions and conditions—not only big type.

63. A material claim that turns one example into typical use

A fictional science fair report shows a battery lasting six hours in one simple circuit. The student says that type of battery always powers any school project for six hours, regardless of load or conditions.

Claim diagnosis: A test result belongs to its setup. Different devices may draw different current; age, temperature and measurement rules may change runtime. The observation does not prove a universal battery property.

Improved wording: “In the reported circuit and test conditions, this battery powered the device for six hours. The result may differ in other loads or environments.”

Try it: A paper plane travels eight metres in one indoor throw. May a student claim the design’s average range is eight metres?

Answer and evidence boundary 63

No. One throw records one outcome, not an average. Repeated trials under documented conditions are needed for a useful distribution or mean.

If the record changes: After repeated controlled trials, an estimated average and variation can be reported for those conditions. The result still does not guarantee identical outdoor performance.

Transfer principle: Experimental nouns and units help attach a result to the conditions that created it.

64. The strongest recommendation drawn from a modest source

A teacher asks for a balanced argument about school libraries. A student finds a credible small study suggesting that guided reading conversations help one group explain complex texts. The conclusion demands compulsory daily discussions for every pupil.

Claim diagnosis: The source may support considering or trialling a strategy, not automatically mandating its use universally. Policy requires feasibility, workload, pupil needs and evidence about wider effects.

Improved wording: “The study offers a reason to consider guided reading discussions for similar learners. A school could test the approach and evaluate outcomes before adopting a universal requirement.”

Try it: A class trial finds that quiet music helped several pupils complete a drawing task. Should a writer demand music for all examinations?

Answer and evidence boundary 64

No. The source concerns one task and a particular group. A defensible proposal might invite further testing of optional background sound in suitable settings, subject to school assessment rules.

If the record changes: A wider evaluation with diverse tasks and student needs may justify more specific recommendations, but the decision remains distinct from the empirical finding.

Transfer principle: Evidence can motivate a proportionate proposal without dictating one inevitable policy.

The practice laboratory: repair an entire paragraph, not just one adjective

Lab 1 — A proposal to change the timetable

The invented school report contains these statements: “Twenty-four volunteers filled in a questionnaire about the morning timetable. Seventeen preferred a later beginning. Teachers did not take part. The questionnaire was shared through one year group’s chat and remained open for two days.” The pupil’s draft says: “Most people in our school want a later start, and the survey proves it would raise achievement.” Before rewriting, underline the phrase “most people” and the verb “proves”. The first expands the population; the second assigns a learning outcome that the questions did not even measure. The questionnaire concerns stated preference among volunteers, not attainment or implementation.

A defensible paragraph might read: “In a voluntary two-day questionnaire shared through one year group’s chat, seventeen of twenty-four respondents preferred a later start. This result suggests that the proposal interested many of those who responded, but the sample does not establish the view of the whole school. The survey also did not measure whether a change would improve learning. A wider consultation involving pupils and staff could help examine practical preferences, while a separate evaluation would be needed to assess academic effects.” The replacement is longer for a reason: it separates a reported preference, a sampling limit, an outcome limit and two different next evidence routes.

Now remove the word “voluntary” from the original source description, without supplying a new recruitment method. Does the writer gain the right to call it random? No. A missing statement about selection is not evidence of randomness. Then change the source to say the poll was a properly documented full-year-group census with twenty-four responses out of thirty invited. The population argument changes, but the missing outcome measurement remains. This is an excellent transfer check because one limitation disappears while another survives. Learners should be able to preserve and revise different parts of the claim independently.

Lab 2 — A science display about plastic waste

An imagined school project weighs plastic waste in a single classroom before and after a poster campaign. The first bag weighs 6 kg; the second weighs 4 kg. The student’s caption says, “Our campaign has reduced all school plastic waste by a third.” The arithmetic difference between 6 and 4 is 2 kg, which is one third of the original 6 kg. Yet the measurement covers one classroom, two selected periods, and perhaps changing student numbers or rubbish collection arrangements. An accurate ratio does not make a claim about the entire institution accurate.

The first repair is: “In the two measured collection periods for this classroom, the recorded plastic-waste mass fell from 6 kg to 4 kg, a reduction of about one third relative to the first measurement.” This is a useful observation. The further sentence “The posters caused the decrease” is not established unless the experiment can distinguish the campaign from other changes. For a school exhibition, the pupil could propose recording more weeks, checking collection consistency and noting whether the number of pupils or events changed. This is a recommendation for better measurement, not a reason to hide the existing result.

A changed-condition challenge: suppose the school discovers that the second bag was collected after one fewer school day. The total masses are no longer directly comparable as rates. The pupil needs to re-evaluate the design, perhaps using mass per comparable day under similar circumstances. Another challenge: suppose the second bag was weighed wet after rain, while the first was dry. The measurements may reflect water as well as plastic. The vocabulary mass, period, comparable, measured, confounded, estimate helps protect a scientific claim from the seduction of a simple headline.

Lab 3 — A Literary interpretation with a narrator’s overstatement

Consider an original invented passage: “When the new pupil stepped into the courtyard, every face seemed to turn away. I heard laughter, though nobody spoke to me.” A student writes, “The writer proves the entire school disliked the newcomer.” The first question should be: Who is speaking? It is a first-person narrator. “Every face seemed” belongs to their experience and perception, not a documented survey of other people’s private feelings. The laughter might be relevant, but without further context it does not establish its target. A close reading should take the narrator’s anxiety seriously without promoting it into universal knowledge.

A better analysis reads: “The narrator feels isolated and interprets the courtyard as unfriendly. The generalising phrase ‘every face’ and the cautious verb ‘seemed’ show a perception coloured by the moment, while the unexplained laughter leaves its cause uncertain.” The sentence can also discuss the effect on readers: they may share the narrator’s discomfort because the scene withholds reassuring detail. Notice that textual criticism can use confident verbs—“the passage presents”, “the wording evokes”—without claiming to know the author’s private beliefs or every character’s motives.

Change one line: a classmate approaches and says, “We were laughing at the runaway kite, not at you.” The interpretation should now incorporate misreading and reversal. The narrator’s first impression remains a real part of the story, yet readers acquire evidence that challenges its accuracy. A Secondary 4 pupil shows control when they can explain both the initial effect and the later revision. A weak answer simply insists the earlier interpretation was “right” or “wrong” without tracing the developing evidence.

Lab 4 — A school-news interview with a missing method

A student preparing a feature interviews five pupils in the library and asks how they feel about silent study spaces. Three want more silent spaces; one prefers group work; one is unsure. Their draft headline is “Students Reject Group Learning”. The phrase reverses what was measured. The interviews asked about preference for silent spaces, not a proposed ban on collaboration. More importantly, five pupils in one location do not represent the whole year group. The proper first move is to preserve the precise question and the five responses.

One faithful account is: “Three of five pupils interviewed in the library said they wanted more silent study spaces, one preferred group work and one was unsure. Their responses illustrate different needs; the interviews do not establish the preference of every pupil.” An editorial can build an argument from those needs: schools might consider providing both quiet and collaborative settings. But the policy proposal belongs to the writer as a recommendation, not to the five interviewees unless they were asked and agreed. A single well-controlled sentence can prevent invented consensus.

Now the adviser asks for a quotation from one interview. The pupil must choose an exact line, keep material conditions intact and seek the speaker’s permission under school procedures. If the interviewee says, “I prefer quiet for revision, but group discussion helps me understand new topics,” quoting only “I prefer quiet” risks concealing the meaningful qualification. A more useful summary preserves the conditional distinction. The source can support different learning conditions for different tasks without declaring one method universally superior.

Lab 5 — A persuasive paragraph that can still have a strong conclusion

Students sometimes fear that limitations will make an essay weak or indecisive. Consider a proposal to provide clearer wayfinding signs around a fictional school campus. A small accessibility walk reveals that two pupils struggled to locate an unfamiliar room. The writer may reasonably recommend testing and improving signs, even though the observation cannot prove every visitor has the same problem. It is possible to make a persuasive recommendation proportional to the practical risk. The evidence identifies a specific potential barrier; the proposal addresses that barrier while inviting better checking.

A strong paragraph could read: “During a supervised wayfinding task, two pupils were unable to find Room C7 from the main entrance without asking for directions. This limited observation cannot establish how often all visitors experience the problem, but it identifies a practical difficulty worth investigating. The school could test clearer directional signs with a wider range of users and review whether the route becomes easier to follow.” The language is neither evasive nor inflated. The writer knows what was observed and what would justify a broader recommendation.

Suppose the next test finds that most participants can locate the room after signage changes, but one wheelchair user cannot reach the route shown. A careless conclusion would celebrate universal success. A more intelligent one would identify the improvement on the tested navigation task alongside a remaining accessibility gap that needs attention. The limitation note should help produce a better decision, not merely protect the writer from criticism. Strong evidence language is a tool for action.

How to practise over four lessons and a delayed check

Lesson one should focus on claims with visible scope words. Give learners five short invented sources and five overconfident conclusions. Ask them to match each source to its actual subject and rewrite one sentence with the right population and time. Avoid introducing sampling theory and rhetorical modality all at once. The teacher should look for the first unstable relation: is the learner misreading a quantifier, confusing an estimate with a measurement, or not realising that the source’s speaker matters?

Lesson two can compare source types. Let pupils explain why a notice is suitable for a schedule, why an eyewitness is useful for a witnessed action, and why neither automatically proves a universal causal claim. The learner may write a two-column chart of “good for” and “not enough for”. This is not a ranking of respectable and disreputable genres. A diary can be extremely useful for studying one person’s experience at a historical moment. A polished institutional summary can still be wrong for a question if it conceals a changed definition.

Lesson three turns reading into writing. Provide a deliberately overstated paragraph and ask pupils to improve it while preserving its useful claim. Mark the first version and the repaired version side by side. The pupil should be able to point to one phrase that changed the population, one that clarified the source, and one that identified a needed next check. A typical weak revision simply adds “possibly” to every sentence; a stronger one decides where firm description is already earned and where limits genuinely matter.

Lesson four practises speaking: pairs exchange claims and ask calm challenges such as “How were the participants chosen?”, “What does the figure count?”, “Does this show a cause or only a sequence?”, and “Could another reading fit the quotation?” The respondent should answer with what is known and identify a next source if needed, rather than treating every question as hostility. A useful oral performance can be concise: “The chart shows a rise in that period, but I cannot tell from it alone what caused the change.”

For a delayed check, provide an unseen task with a new format—perhaps a school poster with a footnote, a small survey, a translated quotation or a fictional graph. Remove the sentence bank and ask the learner to produce a note with literal evidence, limitation, maximum defensible claim and next evidence. Record how many prompts were required. If the child can do the exercise only with a teacher nearby, that is still progress from not understanding the method, but independence has not yet been shown. A follow-up should target the missing retrieval or decision step.

Progress checks that examine reasoning, not performative hesitation

Progress should not be measured by counting the number of times a pupil uses “may” or “perhaps”. Over-qualification can make a clear factual statement less useful. A stronger measure is whether the learner selects a confidence level appropriate to the claim. Can they firmly report what a timetable says? Can they distinguish a survey’s respondents from the full population? Can they refuse to infer cause from a simple time series? Can they revise a conclusion when a footnote or new source changes its scope? These are specific actions that can be observed.

In a short assessment, present four claims of different evidential strength. One source explicitly says a workshop was scheduled for Tuesday; the correct answer should describe its schedule without extravagant hedging. Another reports an unverified online rumour; the answer should attribute and avoid endorsement. A third provides a detailed local observation; the answer should retain its local scope. A fourth offers a well-described comparison under limited experimental conditions; the answer may support a cautiously causal interpretation within those conditions. Such variety stops students from treating caution as a universal sentence template.

The student might improve in one area while struggling in another. They could become excellent at detecting quantifier errors yet continue to confuse correlation and causation. They might write accurate notes untimed but overstate claims when rushing an essay. Record the conditions of the work, including whether a checklist was available, whether sources were supplied, and how much feedback was given. One perfect revision beside a model should not be mistaken for fluent independent judgement across contexts.

The final educational goal is to reduce the need for visible scaffolding. First the learner might use the full four-column ledger. Later they can mark scope directly in the source margin. In an examination-style writing task, they may simply pause for ten seconds and choose the accurate verb. The scaffold has done its job when it supports a durable thinking habit, not when the learner carries a bulky template into every paragraph.

What a parent, a teacher and a student can each do

A student’s useful next action

Choose one paragraph from school practice and circle the most consequential claim word. Write the source’s actual supported fact beneath it. If the two do not match, change the claim or find relevant further evidence. Then explain the revision to a partner in ordinary language. There is no advantage in making the note obscure; a precise sentence understood by another reader is stronger than a paragraph of technical language whose scope nobody can recover.

A parent’s useful next action

Ask what a source actually says before praising how convincing an answer sounds. A parent can read two invented survey cards and ask, “Who answered?”, “When?”, and “What were they asked?” without knowing statistical terminology. Encourage pupils to admit uncertainty accurately. If a child says the answer is “not established”, the next question is “What would help establish it?”, not an automatic accusation of indecision. Academic confidence grows from checking what one can justify.

A teacher’s or tutor’s useful next action

Build a small source pack with deliberately different evidential jobs. Include one exact rule, one voluntary survey, one short interview, one chart with a changed axis, and one fictional narrative excerpt. Ask pupils to identify the strongest defensible statement for each and to explain the boundary. Offer one worked model, one guided task and a new independent task. Feedback should target the error’s location—population, time, outcome, causality, attribution, measurement or quotation—rather than saying the entire argument “lacks depth”.

Frequently asked questions

1. Does every essay need a formal source-limitation paragraph?

No. An appropriate qualification can appear within one clear sentence. A classroom planning note can be useful behind the scenes without appearing as a mandatory extra section in examination writing.

2. Will qualifications make a persuasive argument weak?

Not if they are precise. A recommendation can remain firm while acknowledging that a small study does not establish universal effect. Weakness comes from vague claims, not from proportionate confidence.

3. What is the difference between unreliable and limited?

A source may be reliable for a narrow job yet limited for another. An official event notice may accurately report a schedule but not actual attendance. A claim of unreliability should identify relevant flaws rather than serve as a general insult.

4. When may we write “proves”?

Only when the intended sense and evidence genuinely support that strength, and even then simpler precise language may be preferable. A single classroom observation rarely proves a broad empirical or causal claim.

5. Is one witness enough?

It depends on the question and consequence. One first-hand account may describe one experience. Important disputes about exact times, motives or wider patterns may require corroboration. Never turn one account into the voice of an entire group.

6. Can a photograph provide evidence?

Yes, for visible features of the depicted frame and moment, subject to authenticity and context. It rarely establishes the whole preceding event, a person’s motive or the frequency of an occurrence.

7. Why is a volunteer poll not automatically random?

Random sampling describes how people were selected. A link completed by interested volunteers does not establish random selection even when the respondents themselves are ordinary pupils.

8. Is uncertainty the same as saying a claim is false?

No. “Unresolved” means the available check did not settle the matter. “Contradicted” means relevant evidence supports an opposing finding. The actions a writer takes should reflect that difference.

9. Why do we need source dates?

To distinguish publication, event, observation, update and access times. A historic diary can be appropriate for a past viewpoint while unsuitable for a current timetable.

10. Should the student cite every sentence?

Citation practice depends on task and context. Important factual claims and quotations need traceable support where sources are being used, but readers also need coherent original explanation rather than a chain of unexplained links.

11. Is this an official O-Level English method?

No. The four-column note is an original educational aid for reading and writing honestly. Check SEAB’s year-specific syllabus for actual assessed tasks and formats.

12. How do we measure improvement?

Use unseen claims of varied types, with reduced scaffolding and delayed retrieval. Check whether the pupil distinguishes firm evidence from inference while retaining useful meaning, not whether every sentence contains a cautious word.

Primary references, canonical routes and review note

Reviewed 26 October 2026. The Singapore Examinations and Assessment Board’s 2026 GCE O-Level school-candidate listing identifies English Language syllabus 1184; its component descriptions should be checked for the applicable cohort, not assumed from examples here. The England English programmes of study are a separate system reference for explicit evidence and precise writing, not an equivalence claim between grades or qualifications. The US Institute of Education Sciences secondary writing practice guide provides evidence-oriented instruction ideas; these are not an official Singapore mark scheme.

Use the existing Vocabulary Learning Hub for the full language progression and the Secondary 4 rehearsal incident-note article when the reader’s job is to record a particular event without reconstructing motives. For provisional timing and audience actions use the existing Secondary 4 delay-update specialist. This article owns the narrower task of calibrating a general claim to a source’s limits, not an apex or replacement for either owner.

All classrooms, surveys, figures, names, photographs, diaries, quotations and outcomes in the original teaching cases are fictional. They illustrate how to reason with evidence, not actual student experiences, survey results, products, school rules, psychological diagnoses or educational guarantees. Good writers do not have to know everything. They need to tell readers accurately what is known, what is inferred, what remains unresolved and what further evidence would make a larger claim legitimate.

Four final challenges: decide what to retain when the source changes

Challenge A: the new statistic does not repair the old question

A fictional group first asks ten pupils whether they enjoy art club, and eight say yes. The next month it asks a hundred pupils whether they attended an after-school activity, and seventy say yes. An enthusiastic writer says interest in art club fell from eighty to seventy per cent. This is an invalid comparison before any sampling objection is considered, because the two questions do not measure the same thing. Eighty per cent of one group expressed enjoyment of one club; seventy per cent of a different group reported participation in any activity. Even if the later sample were much larger, it could not repair the difference in construct. More data on a different question is not automatically better evidence for the original question.

A useful revised report names both results separately and asks which comparison is actually needed: “Eight of ten art-club respondents in the first poll said they enjoyed the club. In a later, broader activity survey, seventy of a hundred respondents reported attending at least one after-school activity. Because the questions, groups and time periods differed, the figures cannot establish a trend in art-club enjoyment.” The next research step would ask a consistently worded question about art club among a clearly defined group across relevant times. The student should be able to distinguish a changed sample size, changed question and changed period without collapsing them into “the data is unreliable”.

Challenge B: two sources agree but do not share an observation method

A website tells readers that a fictional river was “extremely clean”, citing a local flyer. A second article uses the identical phrase, citing the website. The student collects both links and announces two independent confirmations of excellent water quality. A source-limitation note should trace the chain backwards: article two repeats article one, and article one repeats the flyer. All three sources may document the same public claim, but none has yet supplied water measurements or stated what “clean” means. The correct response is to treat the repetition as one evidential lineage and inspect whether an appropriate original monitoring result exists.

Now suppose a laboratory report becomes available listing measured indicators with dates and locations. That is stronger evidence of the tested properties during the documented period, but it still may not justify “safe for every activity” or “always clean”. The learner needs the measured variable, criterion and sample context. A river could meet one chemical measure while failing another condition. The exercise teaches that even a seemingly objective result must be linked to a defined claim; the task is not to replace all descriptive English with laboratory jargon. An accessible summary might state, “The report found the tested value below the specified limit at the sampling locations in May,” followed by the relevant limitation.

Challenge C: the quote belongs to a person, not to a category

A fictional interviewee says, “I enjoyed learning through the theatre rehearsal more than through worksheets.” A student writes, “Performers prefer practical learning over written tasks.” The source concerns one person’s preference in a particular context. Turning it into a statement about performers as a whole creates an unsupported group identity and may stereotype people who have diverse learning needs. A faithful use would attribute the preference, perhaps discussing why the rehearsal helped that individual. The writer may still explore a broader question about practical and written learning, but needs more sources for any broader conclusion.

The changed condition is that the interviewee later says, “I still needed written notes to remember the dialogue.” This does not make the earlier enjoyment false; it qualifies any imagined opposition between rehearsal and writing. A sophisticated argument can explain that a person may prefer one mode for engagement and another for retention. The vocabulary preferred, benefited, relied on, complemented, contrasted, under these conditions lets a writer represent a complex account without making it vague. Another pupil who prefers written planning should not be made into a counterexample against a universal claim that was never justified in the first place.

Challenge D: a true figure used for the wrong denominator

A school report says that thirty pupils joined a voluntary debate club and twelve completed an optional advanced workshop. A student writes, “Forty per cent of all Secondary 4 students completed the workshop.” The calculation twelve divided by thirty is forty per cent, but the denominator of thirty covers club members, not necessarily all Secondary 4 pupils. The arithmetic may be correct for the club population while false as a year-group claim. The first revision is to name the base: “Twelve of the thirty debate-club members completed the optional workshop.” This statement must also be checked to ensure the twelve are indeed drawn from those thirty and not a partly different group.

Now another record shows twelve workshop completers, but only ten belonged to the debate club and two were invited from another class. The original denominator link fails. The writer should not continue using forty per cent of club members. If the question asks how many club members completed the workshop, the answer is ten of thirty, or one third, assuming the records cover the same period and no further entries are missing. If the question asks what share of all workshop completers belonged to the club, it is ten of twelve. Both are valid calculations answering different questions. A good source-limitation note names the population that owns each number before choosing the fraction.

One final test links the four challenges. For any unfamiliar document, the learner should identify the speaker or data owner, the population or object, the measured or asserted relationship, and the time and context. Then they should write one accurate sentence and one sentence explaining an important boundary. If a new document changes only the date, keep the remaining sound reasoning. If it changes the denominator, repair all dependent percentages. If it supplies a new quotation, preserve the original quotation’s speaker and context. The point is to teach flexible control: the learner can make a stronger claim when stronger evidence arrives, and narrow an older claim without treating correction as defeat.