First, find what the advisory actually says
A school health advisory may say that a habit increases a risk, that certain conditions make exposure more likely, and that readers should take a practical precaution. These are three different messages. A stronger English comprehension answer does not turn “may increase” into “will cause”, does not confuse a recommendation with an observed result, and does not invent a reason that the text never gives. Read the claim, find its qualification, and answer only the question that was asked.
This G3 English Paper 2 guide follows Vol 0568, on reading unequal breath-count observation windows, and Vol 0567, on distinguishing the effects of tobacco-smoke components. Here the learning goal is not medical diagnosis. It is precise reading: who is addressed, what evidence is actually presented, when an instruction applies, how language creates an effect, and what cannot fairly be concluded. Every advisory and question below is original fictional practice, not an official SEAB examination passage.
Quick route: identify the main claim; circle its condition and qualifying words; separate fact, interpretation and recommendation; map pronouns to referents; explain the author’s purpose using precise evidence; paraphrase without strengthening or weakening the claim; then practise examination-style answers under a short time limit. Keep scientific knowledge in its proper place: general background may help you understand a topic, but it cannot replace what a source text states.
The official 2027 SEC English examination: what is being practised
The SEAB 2027 G3 school-candidate catalogue, updated 26 October 2026, lists English Language as K300 (former O-Level reference 1184). The official K300 syllabus specifies Paper 2 Comprehension, 50 marks, 35% of the English assessment, with a duration of 1 hour 50 minutes. Section A uses two texts, one involving visuals; Section B uses a narrative or recount; Section C uses a non-narrative text and includes a summary of about 80 words. The question types and exact source texts vary; this chapter exercises non-narrative interpretation, not a prediction of a particular health passage.
For the wider reading route, the site’s Secondary English Comprehension owner teaches inference and language effect. Vol 0557 examines tone and language effect in serving-label texts. The earlier PSLE learner guide about holding a definition steady supplies a useful foundation: a word must not change meaning midway through an answer simply because a familiar topic invites guesses.
The first distinction: statement versus instruction
“The room was poorly ventilated” reports a condition. “Open the windows” tells somebody to act. “Opening the windows can improve air exchange” links an action to a possible effect. In a comprehension answer, these sentences cannot be treated as interchangeable. When asked for a precaution, quote or paraphrase the instruction. When asked for the condition that prompted it, find the described situation. When asked for a reason, use the source’s explanation rather than supplying a plausible scientific explanation of your own. One paragraph can contain all three, so classify each clause before answering.
A risk is not a certainty
Risk language concerns likelihood or exposure, not a guarantee about every person. “Can increase the risk” does not mean “always causes”. “May” marks possibility; “is associated with” describes a relationship without, by itself, proving the direction of cause. When you paraphrase these expressions, conserve their strength. The safest answer is not always the vaguest: if the passage makes a definite statement about a measured observation, report it definitely. The task is to match the source’s level of confidence exactly, rather than making every answer cautious by habit.
When the word only changes everything
Compare “only students with permission may enter” with “students may enter only with permission”. These are closely aligned. But “students with permission may only enter” can suggest that entry is their sole permitted activity. Placement affects meaning. The same issue appears with “not all” and “all do not”. “Not all samples met the standard” allows some to pass; “no samples met the standard” allows none. A single word can reverse an inference. In longer questions, underline the scope word and draw a bracket around the phrase it limits.
Conditions and exceptions must travel together
A guideline may recommend an action for a group except under stated circumstances. A full answer takes both the main instruction and the exception. For example, “Move the activity indoors when a weather warning is issued, unless the indoor venue is closed” has a trigger and an exception. The answer “Always move indoors” is wrong because it discards the condition. The answer “Cancel the activity” is wrong because cancellation was not stated. Use an if–then–unless map and check every branch before settling on a response.
Causation is stronger than association
“More people reported discomfort on the day the windows were closed” establishes a coincidence in that record, not that the closed windows caused every symptom. “The trial showed that opening windows reduced the measured gas level in this room” concerns a particular measured variable and setting. English examination questions can ask what a writer concludes, assumes, suggests or admits. Read the verb in the question. “Suggests” does not demand proof, while “demonstrates” calls for stronger textual support. Give an answer with the same level of logical force as the evidence.
Population claims and individual claims
A survey about a group describes that group under its collection conditions. It cannot establish an individual person’s health status. “Most respondents preferred the shorter notice” does not mean “everyone preferred it”. “An average changed” does not mean “each observation changed”. This distinction is central when a passage discusses breathing, exercise or health. Do not diagnose a fictional student from a classroom activity. In an exam, the correct textual inference may be that the writer warns against treating a group summary as a personal prediction.
Read numbers with their denominator
An advisory says that eight out of ten respondents found a notice clear. Before answering, look for the count, denominator, selection method and exact question. Eight out of ten is eighty per cent of those respondents, not eighty per cent of a school unless those ten represent the whole population and the claim says so. A figure in a passage is evidence only for the proposition to which it relates. A seemingly impressive percentage does not automatically validate a larger or different statement in a headline.
Time changes what a comparison means
“Before the session” and “after the session” describe two moments, not necessarily an intervention effect. A text may distinguish a temporary change from a long-term outcome. If an author says “Immediately after exercise, the count was higher”, the reply “exercise permanently raises the count” is unsupported. Trace time adverbs, durations and the order of events. When a paragraph includes “initially”, “later” and “finally”, sketch a three-part sequence. A clear timeline prevents wrong causal claims and helps answer questions about contrast or development.
Audience and purpose govern language choice
An advisory addressed to parents usually differs from a scientific report written for specialists. Words such as “check”, “ask” and “remember” invite action; technical terms may be explained rather than assumed. To identify purpose, connect form and wording: the writer defines a term to make instructions accessible, cites a qualification to stop overconfidence, or places an instruction first because readers need to act quickly. Avoid bland answers like “to attract attention”. Explain which reader and which action or understanding the writer is trying to shape.
A claim and its supporting evidence are separate
A paragraph may open with a conclusion and follow with a count, quotation or example. Label those pieces claim, evidence and qualification. If a question asks for two reasons, do not provide a statistic and a rewording of the same statistic as two reasons. If it asks why evidence is convincing, show how that evidence bears on the point. If a passage provides no evidence for a large claim, you may say so when the question calls for evaluation. You should not supply new evidence from your own memory and attribute it to the writer.
Language effect needs a specific effect
For “a silent warning light”, a good response might explain that “silent” suggests danger can be overlooked because it does not announce itself loudly. Saying “it makes the reader interested” is too general. For an imperative like “Check the conditions before you share”, identify the action urged and the reader’s responsibility. For a question such as “Would you trust a claim without its source?”, explain how the direct question invites readers to examine their own habits. Attach an effect to the actual wording and audience.
A fair paraphrase preserves meaning
Paraphrasing is not merely swapping synonyms. “Not enough information to conclude” cannot become “wrong”; “some evidence” cannot become “strong proof”. Rewrite a sentence by preserving actor, action, limitation, timing and consequence. A useful test is reverse translation: could someone reconstruct the source’s main qualification from your answer? If not, your paraphrase has lost important meaning. Concise answers should trim decoration, not alter logical content. Never write an impressive-sounding sentence that means more than the passage.
Distinguish advice from a rule
“You are encouraged to check” is softer than “you must check”. “Should” indicates recommendation or expectation in many texts, while “must” states necessity; both still depend on the source’s context. In a comprehension answer about authority or urgency, examine modals, not just subject matter. A safety advisory may contain instructions that sound serious, but the exam asks how its wording works. Respect the precise force of a modal instead of automatically replacing every recommendation with a command.
Understand reference chains
The word “this” may refer to a whole preceding sentence, not simply the nearest noun. “The readings were collected at different times. This prevented direct comparison.” Here “this” refers to different collection times. To resolve a pronoun, replace it with your proposed antecedent and reread the sentence. If the result is awkward or logically false, inspect the larger context. Questions about “they”, “such findings” and “the practice” test attention to reference and meaning, not just vocabulary recall.
Contrast markers show the author’s position
“However”, “although”, “nevertheless” and “instead” signal a turn, but the relationships differ. “Although the chart is clear, its sample is small” concedes readability and limits evidential reach. “Instead of making a claim, the school asked for more observations” expresses replacement of one action by another. Do not write “the author disagrees” without naming the proposition under review. A strong answer states both sides and the role of the connective: concession, contrast, correction or alternative.
Headlines and body text may differ in strength
A headline may be compressed to invite reading, while the body explains boundaries. An examination question may invite you to evaluate whether a heading fairly represents the passage. Quote the key difference without accusing the writer of dishonesty unless the evidence supports it. A heading saying “A safer school” can be an aspiration, while a paragraph saying “the proposed practice may reduce one exposure” is much narrower. State the mismatch precisely: certainty, scope, timeframe or subject.
Do not confuse comprehension with subject knowledge
You may know that oxygen diffuses across alveoli and that carbon monoxide can reduce oxygen transport. Those facts help you follow a passage, but they should not be inserted when the comprehension question asks what a writer says about an advisory’s wording. If the text says “ask a trained adult to review the readings”, answer that instruction, not a biology lesson. Where a passage does invoke science, distinguish direct quotation, reasonable inference and outside knowledge. This protects marks and helps keep public copy responsible.
Summary is a separate task with a specific boundary
In K300 Paper 2 Section C, the official syllabus includes a summary of about eighty words, excluding given introductory words. If a health advisory were used in a summary task, you would select the points requested by the task rather than summarise the whole topic. Preserve warnings and exceptions if they are essential points; remove illustrations when they add no distinct information. Draft a list of eligible ideas, combine only genuinely related items, and count approximately rather than forcing unnatural grammar to fit a target.
Choose evidence before you phrase the answer
A common error is to write an attractive sentence and then search for a quotation to support it. Reverse that order. Identify the relevant sentence, locate the exact noun or condition, and decide what it proves. Only then craft the answer. This is especially important when a text contains numerical data, a personal anecdote and a general recommendation. An anecdote may illustrate concern without supporting a statistical conclusion. State what kind of evidence you are using and why it fits.
Finish with a bounded conclusion
A strong final sentence does not overstate what the text establishes. If a school trial recorded a change during one session, write that the result was observed in that session; a larger conclusion would need more data. If a writer recommends checking multiple sources, state that recommendation rather than inventing a policy. A bounded conclusion can be precise and confident. The goal is not to sound uncertain. It is to give the most definite answer that the available words and evidence justify.
Original worked comprehension questions: twenty-eight short decisions
The miniature passages below are fictional. They exercise the reading moves just taught; they are not predictions of SEAB question topics. Read each stimulus before the explanation, cover the model response, and write your own one- or two-sentence answer. These are guided practice cases rather than a replacement for a complete timed paper.
Worked case 1: The notice says, “Some pupils may notice a temporary change after vigorous activity.” What is not justified?
Model response. It is not justified to say that every pupil will experience a lasting change. Why. “Some” limits the group, “may” marks possibility, and “temporary” limits duration. Losing any one changes the claim. Common trap. A tempting answer is “all pupils will change permanently”, which reverses all three limits. Transfer check. In a new passage, identify the exact source phrase that licenses the answer, preserve its scope and timing, and remove any assumption that the writer has not stated.
Worked case 2: A leaflet reads, “Where windows can be opened safely, staff should improve ventilation.” Identify the condition.
Model response. The instruction applies only when it is safe and possible to open the windows. Why. The opening clause governs the recommendation. It is not a decorative aside; safety is part of the writer’s advice. Common trap. Do not reduce the sentence to “staff should always open windows”. Transfer check. In a new passage, identify the exact source phrase that licenses the answer, preserve its scope and timing, and remove any assumption that the writer has not stated.
Worked case 3: A chart lists thirty observations, but only twenty-four have complete times. What can the caption fairly report?
Model response. It can report analyses based on the twenty-four complete timed observations, if that is the stated method. Why. A denominator must describe the records actually used. The six incomplete records still exist but cannot support the timed rate calculation. Common trap. Do not claim that all thirty observations yielded valid rates. Transfer check. In a new passage, identify the exact source phrase that licenses the answer, preserve its scope and timing, and remove any assumption that the writer has not stated.
Worked case 4: The writer says “there was an association”, then “we cannot infer the cause”. Identify the relationship.
Model response. The writer notes a relationship in the data while explicitly rejecting a causal conclusion. Why. The two clauses are compatible; the second narrows what the first allows us to infer. Common trap. Do not paraphrase association as proof that one factor caused the other. Transfer check. In a new passage, identify the exact source phrase that licenses the answer, preserve its scope and timing, and remove any assumption that the writer has not stated.
Worked case 5: A parent says “the readings prove the plan works”; the advisory says “readings were mixed”. Explain the disagreement.
Model response. The parent makes a certain positive claim, while the advisory reports inconsistent results. Why. The source’s wording supplies a boundary: the evidence does not consistently support the claimed success. Common trap. Do not call the parent dishonest; the passage only establishes overconfidence. Transfer check. In a new passage, identify the exact source phrase that licenses the answer, preserve its scope and timing, and remove any assumption that the writer has not stated.
Worked case 6: “Before sharing the chart, check what its axes measure.” What is the writer’s immediate purpose?
Model response. To urge readers to inspect axis labels before interpreting or repeating a numerical claim. Why. The imperative “check” specifies a concrete verification action and “before” makes it a prerequisite. Common trap. “To make the piece exciting” misses both the action and its timing. Transfer check. In a new passage, identify the exact source phrase that licenses the answer, preserve its scope and timing, and remove any assumption that the writer has not stated.
Worked case 7: “It would be premature to call the first trial conclusive.” What does premature mean here?
Model response. It means reaching a final conclusion earlier than the available evidence warrants. Why. The key meaning comes from “first trial” and “conclusive”, not from a dictionary synonym unrelated to evidence. Common trap. Avoid explaining it as the trial having started early in the morning. Transfer check. In a new passage, identify the exact source phrase that licenses the answer, preserve its scope and timing, and remove any assumption that the writer has not stated.
Worked case 8: “The guide offers three options; only the second requires an appointment.” Which option can be used without a required appointment?
Model response. The first and third options do not have that stated requirement. Why. The word “only” restricts the appointment requirement to the second option. Common trap. Do not assume that the first and third never involve appointments in other circumstances. Transfer check. In a new passage, identify the exact source phrase that licenses the answer, preserve its scope and timing, and remove any assumption that the writer has not stated.
Worked case 9: “Not all classes completed the activity.” Does the statement say that no class completed it?
Model response. No. It says at least one class did not complete it, while allowing some classes to have finished. Why. “Not all” and “none” express different logical quantities. Common trap. Do not treat a partial shortfall as universal failure. Transfer check. In a new passage, identify the exact source phrase that licenses the answer, preserve its scope and timing, and remove any assumption that the writer has not stated.
Worked case 10: The passage notes a short-term change and then a longer investigation. Why is the later study mentioned?
Model response. It shows that the first result may not settle whether the pattern persists. Why. The second study relates to duration of evidence; it does not erase the observed first result. Common trap. Do not state that the first measurement must have been false. Transfer check. In a new passage, identify the exact source phrase that licenses the answer, preserve its scope and timing, and remove any assumption that the writer has not stated.
Worked case 11: “This is why the school retained the original log.” What does this refer to after a sentence about conflicting summaries?
Model response. It refers to the conflict between the summaries and the need to check them against the original data. Why. A reference can point to a preceding situation or explanation, not a single noun. Common trap. Do not replace “this” mechanically with the closest object if the sentence stops making sense. Transfer check. In a new passage, identify the exact source phrase that licenses the answer, preserve its scope and timing, and remove any assumption that the writer has not stated.
Worked case 12: “Even a neat-looking figure can conceal missing cases.” Explain the effect of even.
Model response. It challenges the assumption that visual neatness guarantees complete or dependable evidence. Why. “Even” introduces a surprising possibility in contrast to what a reader may expect. Common trap. A response about the chart being beautiful does not explain that contrast. Transfer check. In a new passage, identify the exact source phrase that licenses the answer, preserve its scope and timing, and remove any assumption that the writer has not stated.
Worked case 13: A school report says “eight of ten respondents” were satisfied. Can an answer say “80% of all parents”?
Model response. No; the stated group is ten respondents, not necessarily all parents. Why. The percentage is 80% within the surveyed ten, but its scope cannot be expanded without evidence. Common trap. Do not invent a representative sampling method. Transfer check. In a new passage, identify the exact source phrase that licenses the answer, preserve its scope and timing, and remove any assumption that the writer has not stated.
Worked case 14: “The school suggested a further check rather than a final verdict.” Identify the contrast.
Model response. The writer contrasts continued verification with a premature final judgement. Why. “Rather than” signals a rejected alternative and the preferred next step. Common trap. Avoid writing simply “they had two opinions” without identifying them. Transfer check. In a new passage, identify the exact source phrase that licenses the answer, preserve its scope and timing, and remove any assumption that the writer has not stated.
Worked case 15: A health note states a clear recommendation but gives no survey count. What can you say about its evidence?
Model response. You can report the recommendation; you cannot attribute a numerical survey finding to this text. Why. The question tests whether an answer supplies evidence that is absent from its source. Common trap. It would be wrong to fabricate a percentage to strengthen the writer’s advice. Transfer check. In a new passage, identify the exact source phrase that licenses the answer, preserve its scope and timing, and remove any assumption that the writer has not stated.
Worked case 16: “These figures describe the class, not each pupil.” What does the author caution against?
Model response. The author cautions against applying a group summary as a definite prediction about one pupil. Why. A collective observation does not necessarily hold for every individual. Common trap. Do not reverse it and claim that group figures have no informational value. Transfer check. In a new passage, identify the exact source phrase that licenses the answer, preserve its scope and timing, and remove any assumption that the writer has not stated.
Worked case 17: The headline says “A proven improvement”, but the body says “one trial suggested a benefit”. Is the headline more definite?
Model response. Yes. “Proven” indicates stronger certainty than “suggested” from a single trial. Why. The issue is certainty and evidence, rather than simply a difference in word length. Common trap. A sound answer identifies the exact words responsible for the mismatch. Transfer check. In a new passage, identify the exact source phrase that licenses the answer, preserve its scope and timing, and remove any assumption that the writer has not stated.
Worked case 18: An instruction reads “Complete the check before the event unless the venue changes.” What do readers need to track?
Model response. They need the normal timing requirement and the exception for a changed venue. Why. The exception can modify what action is appropriate; it should remain attached to the instruction. Common trap. Do not paraphrase the check as compulsory at that timing in every possible circumstance. Transfer check. In a new passage, identify the exact source phrase that licenses the answer, preserve its scope and timing, and remove any assumption that the writer has not stated.
Worked case 19: “One useful precaution is to verify the source.” Why is one important?
Model response. It presents source verification as one useful measure, not necessarily the sole available measure. Why. The quantifier resists an exclusive reading. Common trap. “This is the only precaution” strengthens the sentence without justification. Transfer check. In a new passage, identify the exact source phrase that licenses the answer, preserve its scope and timing, and remove any assumption that the writer has not stated.
Worked case 20: “While the volunteers agreed, the organisers wanted further evidence.” What does while do?
Model response. It marks a contrast between the volunteers’ agreement and the organisers’ caution. Why. In this context, “while” means “although”, not necessarily events occurring simultaneously. Common trap. Read the clauses together before selecting the sense of a connective. Transfer check. In a new passage, identify the exact source phrase that licenses the answer, preserve its scope and timing, and remove any assumption that the writer has not stated.
Worked case 21: The text gives two claims: “Counted 15 breaths” and “the pupil was unwell”. Which is supported by the log?
Model response. The log supports a count of fifteen breaths if reliably recorded; it does not, by itself, establish illness. Why. The first is an observation; the second is a health inference requiring different evidence. Common trap. Do not mistake a familiar context for proof of a diagnosis. Transfer check. In a new passage, identify the exact source phrase that licenses the answer, preserve its scope and timing, and remove any assumption that the writer has not stated.
Worked case 22: “If the recording is incomplete, do not calculate a rate.” What is the reason for the condition?
Model response. A rate requires a count and its observation time, so an incomplete time record cannot support the calculation. Why. The mathematical logic explains why the advice is conditional, but the answer should stay close to the passage. Common trap. Do not invent a rate by assuming a one-minute window. Transfer check. In a new passage, identify the exact source phrase that licenses the answer, preserve its scope and timing, and remove any assumption that the writer has not stated.
Worked case 23: An author includes a parent’s concern before a measured result. How might the structure work?
Model response. It introduces a human reason for the inquiry and then moves to evidence that can be checked. Why. The concern may establish context while the measured result addresses a narrower factual question. Common trap. Avoid treating the concern as a substitute for measurement. Transfer check. In a new passage, identify the exact source phrase that licenses the answer, preserve its scope and timing, and remove any assumption that the writer has not stated.
Worked case 24: “The data were consistent across the three observed sessions.” Can a conclusion extend to the whole year?
Model response. Not from that statement alone; it describes only three observed sessions. Why. Consistency within a recorded window does not establish an unobserved period. Common trap. The correct answer can affirm consistency and still respect its boundary. Transfer check. In a new passage, identify the exact source phrase that licenses the answer, preserve its scope and timing, and remove any assumption that the writer has not stated.
Worked case 25: “Staff will review the plan if further evidence contradicts it.” What commitment is made?
Model response. The plan will be reconsidered if later evidence conflicts with it. Why. The “if” clause gives a clear trigger for review, and the future tense states a planned response. Common trap. Do not report that contradictory evidence has already appeared. Transfer check. In a new passage, identify the exact source phrase that licenses the answer, preserve its scope and timing, and remove any assumption that the writer has not stated.
Worked case 26: The passage says “may reduce exposure”; a student writes “eliminates exposure”. How should it be repaired?
Model response. Replace the claim with “may reduce exposure” or an equivalent qualified expression. Why. Both possibility and degree are changed by “eliminates”. Common trap. A paraphrase is not accurate merely because it sounds positive. Transfer check. In a new passage, identify the exact source phrase that licenses the answer, preserve its scope and timing, and remove any assumption that the writer has not stated.
Worked case 27: “Although the report was brief, it named every excluded record.” What quality is emphasised?
Model response. The report’s transparency about excluded records despite its short length. Why. The concession “although” contrasts brevity with sufficient disclosure. Common trap. Do not conclude that every excluded record was later used. Transfer check. In a new passage, identify the exact source phrase that licenses the answer, preserve its scope and timing, and remove any assumption that the writer has not stated.
Worked case 28: A question asks what the author’s final paragraph recommends, but the answer gives the opening survey result. Why is that weak?
Model response. The answer retrieves information from the wrong part and confuses a result with a recommendation. Why. Locate the question’s target function first, then choose evidence from the relevant passage. Common trap. Giving a correct number can still fail if it does not answer the asked question. Transfer check. In a new passage, identify the exact source phrase that licenses the answer, preserve its scope and timing, and remove any assumption that the writer has not stated.
A complete practice text with questions and model marking commentary
Fictional school advisory: “During a four-day classroom observation, pupils recorded whether the ventilation notice was visible when they entered. Twenty-four of thirty intended observations were completed with both a location and a time. The school found that the notice was seen in eighteen of those twenty-four complete records. This finding concerns notice visibility in the sampled sessions, not whether every pupil understood or followed the guidance. Staff therefore recommended checking placement at different entry points. The school will review the layout again after further observations.”
Question 1: What percentage of complete records reported seeing the notice? The model response is 75%, because eighteen divided by twenty-four gives three quarters. The denominator is complete records, not all thirty intended observations. The common error is dividing by thirty without reading the definition of the usable sample. In a comprehension paper, report the numerical relationship only when the question requests it, and retain the sentence’s restriction.
Question 2: What can the writer not infer from the result? The writer cannot infer that every pupil understood or followed the guidance. Seeing a notice is not the same as understanding it or acting on it. A precise answer identifies both the observed variable and the unsupported leap. “The notice does not work” would be equally unjustified, because the source has not tested that conclusion.
Question 3: What is the purpose of the final sentence? It tells readers that the arrangement remains open to review after more evidence is collected. The writer does not describe the existing trial as worthless; instead, the final sentence connects current evidence to a future checkpoint. This is an example of cautious but useful communication.
Question 4: Explain “therefore” in the recommendation. It marks a consequence of the narrower finding: because the records concern visibility at certain places and times, staff recommend checking placement across different entry points. A good response names that reasoning link. Do not answer that “therefore” simply makes the paragraph flow.
Question 5: Write a brief summary of the practical safeguards described. Suitable points include maintaining complete location-and-time records, defining what was actually counted, avoiding claims about understanding or behaviour that were not measured, checking several entry points and reviewing the layout after more observations. An accurate summary groups related points without introducing new safety facts, uses its own wording and respects the requested word limit.
How to practise this skill without reading mechanically
First read the entire non-narrative text once for its argument. On the second pass mark every modal, quantity, exception and temporal phrase that changes the strength or scope of a claim. Then solve literal questions before inference or effect questions; the factual map will make the deeper work more dependable. Answer in full meaning units but not unnecessarily long paragraphs. Each sentence should earn its place by adding an idea demanded by the question.
For revision, write a two-column error log. On the left put the exact overstatement you made, such as “always”, “proves” or “all pupils”. On the right write the source-faithful replacement, such as “may”, “suggests in this sample” or “some pupils”. A third column can record the trigger that caused the error: unfamiliar science, a vivid anecdote, a large percentage or a misleading headline. This creates transferable control rather than memorising one passage.
The next integrated EMS guide will apply these reading decisions to a school evidence briefing, while the next Biology guide returns to respiratory structures. Review the English comprehension owner if inference and language effect remain difficult. The essential examination habit is simple: make your answer as strong as the source allows, but no stronger.
Official references and reader note
Exam code, paper weighting and syllabus requirements were checked against SEAB’s 2027 G3 syllabus catalogue and the K300 2027 English syllabus. The examples in this guide are original teaching materials. They do not reproduce any confidential examination question, predict a future passage or offer personalised medical advice.
