A notice can be grammatical and still mislead a reader
The habitat receiving bay has produced two kinds of records. One describes the route that light follows through a simplified optical window. Another counts how many carton checks are readable, flagged and awaiting manual review. A bulletin now reports that the inventory is ready. Does the bulletin mean every carton is correctly identified? Does it mean all flagged cases have been checked? Does its confident tone hide an important qualification? These are English comprehension questions, not additional calculations.
This guide is a fresh Paper 2 reading clinic. It teaches how to distinguish what a notice states from what the reader infers, compare an early and late message, recognise the effects of specific words, follow a narrative character’s change of mind and write an accurate short summary. The science and mathematics are background context only; those topics have their own specialist volumes.
Quick route: identify the precise question; find the relevant phrase; distinguish assertion, evidence and inference; choose only the detail that answers; paraphrase without changing meaning; and check the final 80-word summary for omissions and unsupported certainty.
The official boundary: English Language K300
The SEAB 2027 G3 index gives English Language code K300 (the earlier O-Level reference code was 1184). The 2027 K300 syllabus sets out Writing Paper 1 (35%), Comprehension Paper 2 (35%), Listening Paper 3 (10%) and Oral Communication Paper 4 (20%). Paper 2 lasts 1 hour 50 minutes and carries 50 marks. Section A has 5 marks for written/visual texts, Section B 20 marks for a narrative or recount, and Section C 25 marks for a non-narrative text, including a summary of about 80 words. The text-based case below is an original practice resource; it is not an actual SEC paper or an official specimen.
The Physics account of the simplified light path belongs to Vol 1285; the probability and review workload belong to Vol 1286. This guide owns the interpretation of notices and reading responses. The Sengkang Secondary English owner supports a broader sequence of reading and language development, while the PSLE transition route helps learners who need a gentler foundation in inference and summary.
Original Text 1: a notice represented in text
INVENTORY CHECK — CURRENT STATUS. Total cartons on register: 120. Cases marked for manual review: 14. Manual review places available this shift: 12. A readable label is not automatically a verified match. The next update will list completed manual checks. Please ask the supervisor for the latest record rather than relying on a photographed notice.
In an actual Section A, a visual text may include layout, colour and images. This exercise keeps everything in words, so a candidate must not claim a colour or an image that is absent. Instead, observe what is prominent in the wording: the headline establishes status, counts occupy short lines, and a warning qualifies the meaning of a readable label. The invitation to consult the supervisor signals that the notice may be superseded.
Original Text 2: an earlier announcement
EARLY SHIFT BRIEFING. The inventory system expects an average of about twelve manual-review flags from 120 checks, assuming a ten-per-cent flag model. Two reviewers have capacity for twelve reviews in a thirty-minute period if each takes five minutes per case. This is a planning projection, not a report of how many cases have actually been flagged. The team will publish an observed count after checking the register.
Text 1 records fourteen actual flags at a later stage. Text 2 gives an expectation of twelve from an earlier model. Their numbers differ but the texts do not logically contradict each other: one describes a realised count, the other a prediction. A comprehension answer must identify that difference before making a judgement. The purpose of a cautious briefing is often to prepare readers for uncertainty rather than conceal an error.
Original Text 3: a narrative of changing judgement
Mira had always found the green tick reassuring. At the top of the register, a neat row of marks seemed to promise that every task was moving towards completion. She had spent the morning carrying empty trays from the demonstration area, and she was glad to see the word ‘readable’ beside most of the carton records. It was the sort of word that looked like progress.
When the shift lead asked her to draft a short update for the class exhibition, she wrote a sentence quickly: ‘Every carton has been confirmed by the new system.’ She chose the word confirmed because it sounded calm. She had not examined the meaning of the code in the register. For a moment, she imagined the satisfaction of placing the bulletin on the board before lunch.
Her classmate Dev read the draft. He did not cross it out at once. Instead, he pointed to the legend at the bottom of the record. ‘Readable means the label was detected,’ he said. ‘It doesn’t say that the label matches the carton.’ Mira frowned. She remembered the teacher’s warning about confident language, but she had thought it applied mainly to persuasive speeches.
They walked to the notice board and compared it with the earlier plan. There had been twelve reviews in the plan, yet the current register showed fourteen flagged records. Mira initially assumed someone had made an error. Dev reminded her that twelve was an expected count under a model, not a promise. The two extra cases did not prove the model was wrong; they showed the need to report what had actually happened.
Mira returned to her desk. This time she wrote: ‘The current register lists fourteen cartons for manual review. The team has capacity for twelve reviews during the scheduled period, so two cases need a revised review plan.’ The sentence lacked the shine of her first headline. It was also more useful. Anyone reading it could see which conclusion depended on the record and which depended on the time plan.
She placed the new bulletin beside the old one, not over it. The dates were visible. Later, when a visitor asked why the two documents used different numbers, she answered without embarrassment. ‘The earlier one estimated a possible workload. The new one counts cases that were actually flagged.’ The visitor nodded. Mira had learned that clarity could do work that certainty could not.
Original Text 4: a non-narrative explanation of traceable records
A register becomes useful when readers know what each status word means. A ‘readable’ result may indicate that a label’s marks were detected. This is distinct from verifying that the detected identifier belongs to the physical carton. A ‘flagged’ result may require further investigation; it does not, by itself, mean that a carton contains the wrong goods. Verification combines a record with an agreed reference and a defined procedure.
A sound update therefore separates planned quantities from observed quantities. A probability model can help estimate workload. For example, a ten-per-cent flag assumption across 120 independent checks gives an expected count of twelve flags. That expected count is not an observation. A later register containing fourteen flagged entries is an observation about the register at that time. A writer should not replace the observed count with the earlier prediction simply because the predicted number was easier to manage.
Time capacity also requires a clearly stated unit. Suppose each manual review takes five minutes and two reviewers have thirty minutes each. The model gives twelve complete reviews if both reviewers can work independently without interruption. Such a calculation tells the team whether the recorded queue fits that model. It does not identify why a particular carton was flagged, prove an optical defect or authorise a real shipment.
A responsible bulletin retains its sources. The register date should be visible; the meaning of each code should be provided; and any unresolved records should be named as unresolved rather than silently omitted. A photograph of an old notice might continue circulating after the underlying register has changed. For this reason, a public message can direct readers to the most recently verified status rather than suggesting that one page will remain current indefinitely.
Language can narrow or broaden a claim. ‘Fourteen entries are flagged in the current register’ is narrow and testable. ‘Fourteen cartons are defective’ is broader and may not be supported. ‘Two reviews remain outside this shift’s ideal capacity’ describes a scheduling consequence. ‘The system is unsafe’ claims a safety judgement requiring much more information. Careful writing refuses to make those leaps merely to sound decisive.
Readers also need a simple way to correct earlier information. An updated note should identify which part changed and why. The earlier projection need not be treated as dishonest simply because the observed figure differed. A useful correction might read: ‘Our earlier bulletin estimated twelve flagged cases; the present register reports fourteen. We have updated the review plan to reflect the observed count.’ This language preserves history, makes the change comprehensible and lets the reader judge the evidence.
Learn the three levels of comprehension
Literal meaning: what appears
A literal question asks you to retrieve a fact that the text supplies. If the current register says fourteen cases are flagged, write fourteen, not the expected twelve from the earlier briefing. Copy the relevant qualifier if it affects the meaning: ‘current register’ distinguishes the observation from the model. Concision is a strength, provided the answer includes everything asked.
Practise this skill with the four original texts above. Underline the phrase that supports the response, then write the smallest complete answer. If a detail is not recoverable from the passage or a reasonable inference from it, do not quietly import outside knowledge. The examination tests your reading of the supplied text, not the ability to sound authoritative about the fictional habitat.
Inference: what the evidence permits
An inference goes beyond a quoted sentence but must remain anchored in it. From Mira changing the bulletin after reading the legend, we may infer she becomes more attentive to precise definitions. We cannot infer that she has become expert in the scanner’s hardware. Good inference combines a textual clue with a proportionate interpretation. If your idea could apply to any story, it is probably too generic.
Practise this skill with the four original texts above. Underline the phrase that supports the response, then write the smallest complete answer. If a detail is not recoverable from the passage or a reasonable inference from it, do not quietly import outside knowledge. The examination tests your reading of the supplied text, not the ability to sound authoritative about the fictional habitat.
Evaluation: what the author’s choice does
The phrase ‘the shine of her first headline’ suggests the original statement was attractive but superficial. Its contrast with the revised sentence makes the reader value usefulness over confident presentation. A strong language-effect answer names the word or image, explains the association it creates, and connects the effect to this narrative moment. ‘It makes the reader interested’ is rarely sufficient because it does not explain how.
Practise this skill with the four original texts above. Underline the phrase that supports the response, then write the smallest complete answer. If a detail is not recoverable from the passage or a reasonable inference from it, do not quietly import outside knowledge. The examination tests your reading of the supplied text, not the ability to sound authoritative about the fictional habitat.
Source, chronology and comparison
Text 2 is earlier and predictive; Text 1 is later and observational. Those properties are more important than which text has the larger number. Questions that ask whether accounts agree may require ‘yes, in different ways’ rather than a flat contradiction. Date, intended purpose and vocabulary are evidence about the document’s authority for a particular claim.
Practise this skill with the four original texts above. Underline the phrase that supports the response, then write the smallest complete answer. If a detail is not recoverable from the passage or a reasonable inference from it, do not quietly import outside knowledge. The examination tests your reading of the supplied text, not the ability to sound authoritative about the fictional habitat.
A visual-text question without invented visuals
In a paper-based mock that reproduces only text, describe the heading, counts, notice hierarchy and warning. Do not fabricate colours, logos or illustrations that the material does not contain. When an actual examination visual is available, inspect its layout and the relationship between graphics and words. A visual-text response should connect the design detail to the communication purpose.
Practise this skill with the four original texts above. Underline the phrase that supports the response, then write the smallest complete answer. If a detail is not recoverable from the passage or a reasonable inference from it, do not quietly import outside knowledge. The examination tests your reading of the supplied text, not the ability to sound authoritative about the fictional habitat.
Summary is selection, not compression alone
A good 80-word summary begins by identifying the precise task. If asked for steps that make an update reliable, select definitions, separation of prediction from observation, verification, source dates, unresolved items and correction methods. Details about Mira’s feelings are irrelevant even if memorable. A shorter passage containing only major points is better than a complete but unfocused retelling.
Practise this skill with the four original texts above. Underline the phrase that supports the response, then write the smallest complete answer. If a detail is not recoverable from the passage or a reasonable inference from it, do not quietly import outside knowledge. The examination tests your reading of the supplied text, not the ability to sound authoritative about the fictional habitat.
Paraphrase without changing certainty
If the source says ‘may require further checking,’ rewriting it as ‘proves an error’ changes meaning. Replace vocabulary only when the replacement preserves the scope: ‘may require’ could become ‘could need’, but ‘always requires’ is too strong. A summary must remain loyal to the text, not merely employ synonyms.
Practise this skill with the four original texts above. Underline the phrase that supports the response, then write the smallest complete answer. If a detail is not recoverable from the passage or a reasonable inference from it, do not quietly import outside knowledge. The examination tests your reading of the supplied text, not the ability to sound authoritative about the fictional habitat.
Pronoun and reference tracking
When ‘it’ appears in a complex passage, identify the nearest grammatically and logically suitable noun. In ‘The bulletin was updated because it used the wrong count,’ ‘it’ refers to the bulletin, not the register. A reference mistake can reverse a causal relation. Work backwards to the earlier sentence when a pronoun seems ambiguous.
Practise this skill with the four original texts above. Underline the phrase that supports the response, then write the smallest complete answer. If a detail is not recoverable from the passage or a reasonable inference from it, do not quietly import outside knowledge. The examination tests your reading of the supplied text, not the ability to sound authoritative about the fictional habitat.
Distinguish a reason from a consequence
The expected twelve reviews comes from a percentage model. The inability to finish fourteen observed flags in the model shift follows from the time-capacity calculation. A consequence is not the same as the cause of the initial flag count. In a ‘why’ answer, give the step that directly explains the event asked about.
Practise this skill with the four original texts above. Underline the phrase that supports the response, then write the smallest complete answer. If a detail is not recoverable from the passage or a reasonable inference from it, do not quietly import outside knowledge. The examination tests your reading of the supplied text, not the ability to sound authoritative about the fictional habitat.
Writers can correct without denying uncertainty
The narrative lets Mira change her view without portraying every previous plan as foolish. She learns to label different kinds of figures. A thoughtful answer notices that the corrective action involves checking the legend, comparing documents and revising the claim. This is more specific than saying ‘she learns to be honest’, although integrity is part of the interpretation.
Practise this skill with the four original texts above. Underline the phrase that supports the response, then write the smallest complete answer. If a detail is not recoverable from the passage or a reasonable inference from it, do not quietly import outside knowledge. The examination tests your reading of the supplied text, not the ability to sound authoritative about the fictional habitat.
Manage the clock without guessing marks
Paper 2 lasts 110 minutes for 50 marks, but some questions need more reading than others. Reserve enough time to read passages carefully, answer the extended texts and check the summary. Do not treat an arithmetic minute-per-mark ratio as an official instruction. A calm plan includes a few minutes for correcting a mistaken reference and checking that summary wording remains grammatical.
Practise this skill with the four original texts above. Underline the phrase that supports the response, then write the smallest complete answer. If a detail is not recoverable from the passage or a reasonable inference from it, do not quietly import outside knowledge. The examination tests your reading of the supplied text, not the ability to sound authoritative about the fictional habitat.
What an evidence-based final sentence looks like
A conclusion should be proportional: ‘The notice distinguishes a projected workload from the register’s current count’ is well supported. ‘The inventory is now guaranteed accurate’ is not. With practice, readers stop treating confidence as an argument and begin seeing exactly which words in the text justify a judgement.
Practise this skill with the four original texts above. Underline the phrase that supports the response, then write the smallest complete answer. If a detail is not recoverable from the passage or a reasonable inference from it, do not quietly import outside knowledge. The examination tests your reading of the supplied text, not the ability to sound authoritative about the fictional habitat.
Thirty-six worked response clinics
1. Literal count
Question: How many cases are flagged in Text 1? Answer: Fourteen. The word ‘current’ identifies the figure as an observed register count.
Why it works: Do not substitute the earlier expected twelve. The candidate’s task is to match the response to the exact wording of the question. Retrieve when retrieval is asked; infer only when the clue supports inference; explain language effect through a specific choice. In all three cases, keep answer length proportional to the marks and avoid a broad statement that outruns the supplied passage.
2. Planning count
Question: What number does Text 2 project? Answer: About twelve flagged cases under the stated ten-per-cent model.
Why it works: Name projection so the reader knows it is not a later observation. The candidate’s task is to match the response to the exact wording of the question. Retrieve when retrieval is asked; infer only when the clue supports inference; explain language effect through a specific choice. In all three cases, keep answer length proportional to the marks and avoid a broad statement that outruns the supplied passage.
3. Two-minute source check
Question: Why might the figures differ? Answer: An estimate and an observed count describe different things; realised results need not equal expected results.
Why it works: Avoid alleging negligence without text evidence. The candidate’s task is to match the response to the exact wording of the question. Retrieve when retrieval is asked; infer only when the clue supports inference; explain language effect through a specific choice. In all three cases, keep answer length proportional to the marks and avoid a broad statement that outruns the supplied passage.
4. What readable means
Question: Does readable mean confirmed? Answer: No. Text 4 distinguishes detecting a label from checking its correspondence to a physical carton.
Why it works: Use the definition supplied by the source. The candidate’s task is to match the response to the exact wording of the question. Retrieve when retrieval is asked; infer only when the clue supports inference; explain language effect through a specific choice. In all three cases, keep answer length proportional to the marks and avoid a broad statement that outruns the supplied passage.
5. What a flag means
Question: Is a flag proof of a wrong carton? Answer: No. It signals a case for investigation, not a verified mismatch.
Why it works: ‘Suggests checking’ is safer than ‘proves defect’. The candidate’s task is to match the response to the exact wording of the question. Retrieve when retrieval is asked; infer only when the clue supports inference; explain language effect through a specific choice. In all three cases, keep answer length proportional to the marks and avoid a broad statement that outruns the supplied passage.
6. Notice purpose
Question: What is Text 1 trying to achieve? Answer: To provide a current status and direct readers to verified updates.
Why it works: Purpose explains why counts and cautions sit together. The candidate’s task is to match the response to the exact wording of the question. Retrieve when retrieval is asked; infer only when the clue supports inference; explain language effect through a specific choice. In all three cases, keep answer length proportional to the marks and avoid a broad statement that outruns the supplied passage.
7. Timing comparison
Question: Which notice is later? Answer: Text 1 records current flags; Text 2 is explicitly an early projection.
Why it works: Do not assume that numerical size decides chronology. The candidate’s task is to match the response to the exact wording of the question. Retrieve when retrieval is asked; infer only when the clue supports inference; explain language effect through a specific choice. In all three cases, keep answer length proportional to the marks and avoid a broad statement that outruns the supplied passage.
8. Capacity in the notice
Question: What is twelve in Text 1? Answer: The number of available manual-review places during this shift.
Why it works: It is not the register’s total flagged count. The candidate’s task is to match the response to the exact wording of the question. Retrieve when retrieval is asked; infer only when the clue supports inference; explain language effect through a specific choice. In all three cases, keep answer length proportional to the marks and avoid a broad statement that outruns the supplied passage.
9. Opening sentence effect
Question: What does the green tick signal to Mira? Answer: It reassures her and initially encourages her to believe the work is complete.
Why it works: Combine the symbol with her reaction. The candidate’s task is to match the response to the exact wording of the question. Retrieve when retrieval is asked; infer only when the clue supports inference; explain language effect through a specific choice. In all three cases, keep answer length proportional to the marks and avoid a broad statement that outruns the supplied passage.
10. ‘Readable’ as progress
Question: Why does the word attract Mira? Answer: It appears to suggest successful completion before she understands its narrow technical meaning.
Why it works: Explain the gap between impression and definition. The candidate’s task is to match the response to the exact wording of the question. Retrieve when retrieval is asked; infer only when the clue supports inference; explain language effect through a specific choice. In all three cases, keep answer length proportional to the marks and avoid a broad statement that outruns the supplied passage.
11. The quick first draft
Question: What does Mira’s early wording show? Answer: She values a calm, finished-sounding bulletin and has not checked the legend.
Why it works: Do not say she deliberately deceives the visitor. The candidate’s task is to match the response to the exact wording of the question. Retrieve when retrieval is asked; infer only when the clue supports inference; explain language effect through a specific choice. In all three cases, keep answer length proportional to the marks and avoid a broad statement that outruns the supplied passage.
12. Dev’s method
Question: How does Dev challenge the statement? Answer: He points Mira to the legend rather than relying on disagreement alone.
Why it works: Identify the concrete evidence move. The candidate’s task is to match the response to the exact wording of the question. Retrieve when retrieval is asked; infer only when the clue supports inference; explain language effect through a specific choice. In all three cases, keep answer length proportional to the marks and avoid a broad statement that outruns the supplied passage.
13. Mira’s frown
Question: What may it imply? Answer: She is unsettled by a distinction that complicates her confident draft.
Why it works: A frown alone does not prove anger or dishonesty. The candidate’s task is to match the response to the exact wording of the question. Retrieve when retrieval is asked; infer only when the clue supports inference; explain language effect through a specific choice. In all three cases, keep answer length proportional to the marks and avoid a broad statement that outruns the supplied passage.
14. Earlier teaching
Question: Which lesson does Mira recall? Answer: A teacher had warned the class about confident language.
Why it works: Mention the recollection rather than an invented warning about scanners. The candidate’s task is to match the response to the exact wording of the question. Retrieve when retrieval is asked; infer only when the clue supports inference; explain language effect through a specific choice. In all three cases, keep answer length proportional to the marks and avoid a broad statement that outruns the supplied passage.
15. Mathematical misunderstanding
Question: What did Mira first suspect about twelve and fourteen? Answer: She thought there must have been an error.
Why it works: The corrected interpretation is that one figure was expected and the other observed. The candidate’s task is to match the response to the exact wording of the question. Retrieve when retrieval is asked; infer only when the clue supports inference; explain language effect through a specific choice. In all three cases, keep answer length proportional to the marks and avoid a broad statement that outruns the supplied passage.
16. Turning point
Question: Which event leads to revision? Answer: The legend and comparison of the two records change her understanding.
Why it works: A causal answer should name the source check. The candidate’s task is to match the response to the exact wording of the question. Retrieve when retrieval is asked; infer only when the clue supports inference; explain language effect through a specific choice. In all three cases, keep answer length proportional to the marks and avoid a broad statement that outruns the supplied passage.
17. ‘Shine’ as language choice
Question: Why compare the first headline with shine? Answer: The image suggests attractive surface confidence while questioning its substance.
Why it works: State the contrast with usefulness. The candidate’s task is to match the response to the exact wording of the question. Retrieve when retrieval is asked; infer only when the clue supports inference; explain language effect through a specific choice. In all three cases, keep answer length proportional to the marks and avoid a broad statement that outruns the supplied passage.
18. ‘More useful’
Question: Why is the second bulletin better? Answer: It states the observed queue, the ideal capacity and the remaining work.
Why it works: Concrete information lets readers act or ask informed questions. The candidate’s task is to match the response to the exact wording of the question. Retrieve when retrieval is asked; infer only when the clue supports inference; explain language effect through a specific choice. In all three cases, keep answer length proportional to the marks and avoid a broad statement that outruns the supplied passage.
19. Visitor’s question
Question: Why does Mira answer without embarrassment? Answer: She can now explain why two figures legitimately describe different stages.
Why it works: It need not imply that the old document was perfect. The candidate’s task is to match the response to the exact wording of the question. Retrieve when retrieval is asked; infer only when the clue supports inference; explain language effect through a specific choice. In all three cases, keep answer length proportional to the marks and avoid a broad statement that outruns the supplied passage.
20. Narrative theme
Question: What idea does the narrative develop? Answer: Responsible communication distinguishes evidence from reassurance.
Why it works: Use a particular incident as support. The candidate’s task is to match the response to the exact wording of the question. Retrieve when retrieval is asked; infer only when the clue supports inference; explain language effect through a specific choice. In all three cases, keep answer length proportional to the marks and avoid a broad statement that outruns the supplied passage.
21. Text 4’s definition
Question: Why is verification more than readability? Answer: It needs comparison with an agreed reference and a defined procedure.
Why it works: Do not replace the definition with a personal guess. The candidate’s task is to match the response to the exact wording of the question. Retrieve when retrieval is asked; infer only when the clue supports inference; explain language effect through a specific choice. In all three cases, keep answer length proportional to the marks and avoid a broad statement that outruns the supplied passage.
22. Planning and observation
Question: Which one is a prediction? Answer: Ten per cent of 120 gives an expected twelve flags.
Why it works: A statistical forecast is not a record of a specific event. The candidate’s task is to match the response to the exact wording of the question. Retrieve when retrieval is asked; infer only when the clue supports inference; explain language effect through a specific choice. In all three cases, keep answer length proportional to the marks and avoid a broad statement that outruns the supplied passage.
23. Risk of copied photos
Question: Why can photographs mislead? Answer: They can circulate after the live register has changed.
Why it works: This explains the recommendation to consult a later verified status. The candidate’s task is to match the response to the exact wording of the question. Retrieve when retrieval is asked; infer only when the clue supports inference; explain language effect through a specific choice. In all three cases, keep answer length proportional to the marks and avoid a broad statement that outruns the supplied passage.
24. Qualifier in an update
Question: What does ‘under the stated five-minute rate’ achieve? Answer: It limits the capacity claim to a particular model.
Why it works: Without the qualifier, readers may mistake the figure for a guarantee. The candidate’s task is to match the response to the exact wording of the question. Retrieve when retrieval is asked; infer only when the clue supports inference; explain language effect through a specific choice. In all three cases, keep answer length proportional to the marks and avoid a broad statement that outruns the supplied passage.
25. A misleading rewrite
Question: Why is ‘fourteen cartons are defective’ too strong? Answer: Flagged and verified defective are not equivalent statuses.
Why it works: Identify the exact semantic leap. The candidate’s task is to match the response to the exact wording of the question. Retrieve when retrieval is asked; infer only when the clue supports inference; explain language effect through a specific choice. In all three cases, keep answer length proportional to the marks and avoid a broad statement that outruns the supplied passage.
26. Reasonable recommendation
Question: What should a cautious bulletin do? Answer: Preserve the source date, describe unresolved cases and say when records change.
Why it works: This is based on Text 4’s recommendations. The candidate’s task is to match the response to the exact wording of the question. Retrieve when retrieval is asked; infer only when the clue supports inference; explain language effect through a specific choice. In all three cases, keep answer length proportional to the marks and avoid a broad statement that outruns the supplied passage.
27. Evidence and trust
Question: Why keep the earlier notice visible? Answer: The timeline allows readers to distinguish a projection from its later observation.
Why it works: Do not infer that retaining every old notice is a universal rule. The candidate’s task is to match the response to the exact wording of the question. Retrieve when retrieval is asked; infer only when the clue supports inference; explain language effect through a specific choice. In all three cases, keep answer length proportional to the marks and avoid a broad statement that outruns the supplied passage.
28. Vocabulary: projection
Question: Give ‘projection’ in this context. Answer: An estimate of future or possible results from a model.
Why it works: It is not a promise. The candidate’s task is to match the response to the exact wording of the question. Retrieve when retrieval is asked; infer only when the clue supports inference; explain language effect through a specific choice. In all three cases, keep answer length proportional to the marks and avoid a broad statement that outruns the supplied passage.
29. Vocabulary: unresolved
Question: What does ‘unresolved’ mean here? Answer: Not yet settled or confirmed by the review process.
Why it works: Do not interpret it as a failed case. The candidate’s task is to match the response to the exact wording of the question. Retrieve when retrieval is asked; infer only when the clue supports inference; explain language effect through a specific choice. In all three cases, keep answer length proportional to the marks and avoid a broad statement that outruns the supplied passage.
30. Vocabulary: traceable
Question: What does traceable mean? Answer: Linked to identifiable sources and steps so that a claim can be checked.
Why it works: It is about records, not merely geographical tracking. The candidate’s task is to match the response to the exact wording of the question. Retrieve when retrieval is asked; infer only when the clue supports inference; explain language effect through a specific choice. In all three cases, keep answer length proportional to the marks and avoid a broad statement that outruns the supplied passage.
31. Inference restraint
Question: Does the passage show the new scanner is defective? Answer: No. It provides no direct test establishing a malfunction.
Why it works: Reason from evidence, not from the topic’s atmosphere. The candidate’s task is to match the response to the exact wording of the question. Retrieve when retrieval is asked; infer only when the clue supports inference; explain language effect through a specific choice. In all three cases, keep answer length proportional to the marks and avoid a broad statement that outruns the supplied passage.
32. Possible bias
Question: Why might a notice that says ‘all confirmed’ be unreliable? Answer: It ignores the limited meaning of readable and the unresolved queue.
Why it works: A critique should identify the source wording. The candidate’s task is to match the response to the exact wording of the question. Retrieve when retrieval is asked; infer only when the clue supports inference; explain language effect through a specific choice. In all three cases, keep answer length proportional to the marks and avoid a broad statement that outruns the supplied passage.
33. Summary focus
Question: If asked for trustworthy update practices, should you include Mira’s excitement? Answer: No; it is not a procedure or criterion.
Why it works: Selection comes before paraphrasing. The candidate’s task is to match the response to the exact wording of the question. Retrieve when retrieval is asked; infer only when the clue supports inference; explain language effect through a specific choice. In all three cases, keep answer length proportional to the marks and avoid a broad statement that outruns the supplied passage.
34. Summary compression
Question: Can ‘Keep dates, explain each status, record unresolved cases’ replace three longer sentences? Answer: Yes, when the original meaning is preserved.
Why it works: Avoid making a new blanket rule. The candidate’s task is to match the response to the exact wording of the question. Retrieve when retrieval is asked; infer only when the clue supports inference; explain language effect through a specific choice. In all three cases, keep answer length proportional to the marks and avoid a broad statement that outruns the supplied passage.
35. Final cross-check
Question: Does the summary use only information in Text 4? Answer: Check each phrase against the original before submitting.
Why it works: Outside knowledge is not a replacement for close reading. The candidate’s task is to match the response to the exact wording of the question. Retrieve when retrieval is asked; infer only when the clue supports inference; explain language effect through a specific choice. In all three cases, keep answer length proportional to the marks and avoid a broad statement that outruns the supplied passage.
36. Strong closing judgement
Question: Which notice better reports the current state? Answer: Text 1, because it gives the observed fourteen and current review capacity, while Text 2 is only a forecast.
Why it works: The evidence supports a contextual comparison, not a character attack. The candidate’s task is to match the response to the exact wording of the question. Retrieve when retrieval is asked; infer only when the clue supports inference; explain language effect through a specific choice. In all three cases, keep answer length proportional to the marks and avoid a broad statement that outruns the supplied passage.
Original summary task with a model
Task: Using only Text 4, summarise the practices that make an inventory status update reliable. Aim for about 80 words, excluding any introductory words a real question may provide. First make a short selection list: define status terms; verify against references; separate predictions from observations; name time and staffing assumptions; retain dated sources; identify unresolved entries; provide updates and correct earlier figures. Omit the science explanation of refraction and the fictional character’s feelings because neither belongs to the specified summary.
Model summary: A reliable update defines each status clearly and distinguishes detected labels from verified matches. It reports observed counts separately from forecasts and explains the assumptions behind staffing capacity. Writers identify their source dates, preserve unresolved records and avoid treating flagged cases as proven defects. Readers should be directed to the latest verified record when earlier notices may be outdated. Corrections should name what changed and why, allowing people to compare claims with the underlying evidence rather than relying on reassuring headlines.
Check the model against Text 4. It condenses ideas without asserting that every flagged item is wrong, without manufacturing a cause for the extra cases, and without presenting a staffing calculation as a safety judgement. A summary of about 80 words is not a licence to omit important limitations. The reader should be able to reconstruct the source’s major recommendations from the summary even though many examples have been removed.
A practical marking and revision routine
After a timed attempt, highlight each part of the question that needs answering. In one colour on your own worksheet, underline evidence in the source; in another, circle your own inference. Check whether every inference is justified by a nearby clue. For language-effect answers, insist on a specific expression and its effect. For the summary, label each selected source point, count the words and read once for grammatical accuracy. Do not practise by memorising the model’s exact phrasing. Instead, explain why each clause survived the selection process.
The next revision session should change the situation. Replace the carton notice with a school laboratory, transport register or library catalogue, while keeping the same demands: define statuses, read dates, compare claims and compress accurately. If the learner can transfer the reasoning to unfamiliar texts, the skill is becoming dependable. If the answer still depends on memorised phrases, return to one short passage and identify the exact point where the meaning was stretched.
Continue through the subject owners
Read Vol 1285 for physical optics, Vol 1286 for probability and capacity, the Secondary English Sengkang owner for language skills, the Sengkang PSLE preparation route for earlier transition work, and the official K300 syllabus for the current assessment format. The next integrated case will ask which inventory claims survive scrutiny across the three subjects; it is a study exercise, not an additional SEC subject.
