2027 G3 SEC English Language K300 | Learner’s Guide Vol 0185. A sentence can contain accurate facts and still claim more than those facts establish. Compare “the results suggest that the change helped” with “the results prove that the change helped”. The first leaves room for uncertainty; the second presents a decisive conclusion. Learning to hear that difference gives students a more exact way to read comprehension passages, examine persuasive language and write responses that stay close to a source.
This guide begins with one verb and works towards full passages and exam-style explanations. It does not ask students to make every sentence hesitant. It asks them to match the certainty of their wording to the evidence in front of them. All passages and exercises are original practice material, not SEAB papers or official answer keys.
1. Where this belongs in the 2027 examination
The official SEAB 2027 G3 syllabus directory, updated 26 October 2026, identifies English Language as K300. The K300 syllabus retains four components: Writing (Paper 1, 70 marks, 35%, 1 hour 50 minutes), Comprehension (Paper 2, 50 marks, 35%, 1 hour 50 minutes), Listening (Paper 3, 30 marks, 10%, about 45 minutes) and Oral Communication (Paper 4, 30 marks, 20%, about 20 minutes including 10 minutes of preparation). Students sit all four.
Paper 2 includes questions on literal, inferential and evaluative understanding, vocabulary in context and the use of language for effect. Section A includes visual-text comprehension; Section B uses a narrative or recount; Section C uses a non-narrative passage and an approximately 80-word summary. Selecting an evidence verb is a useful learning method across these tasks, though the syllabus does not promise a question explicitly called “evidence verbs”.
Paper 1 asks writers to communicate purpose, audience and context accurately. In Situational Writing, the specified length is 250–350 words; in Continuous Writing, 350–500. The purpose of this chapter is to improve the judgement that controls an answer or claim, not to prescribe a formula for every piece of writing. When practising editing, remember the official Section A tests grammatical errors, not spelling or punctuation errors.
2. Read one small report before answering
Imagine a school opens its library thirty minutes earlier for two weeks. An original report says: “During the trial, average morning attendance rose from 40 visitors to 54. The two weeks coincided with a reading challenge. Staff also noticed that some pupils visited to complete homework. The team has not compared attendance with a similar period without a challenge.”
Question: Does this passage prove that opening earlier caused the rise? No. It shows that the average recorded attendance was higher during the trial. It suggests that the change may have made visiting easier, but other simultaneous changes may have contributed. It does not isolate the effect of opening hours.
The difference is subtle but important. A learner might calculate 54 − 40 = 14 correctly and still answer the causal question badly. “Attendance increased by 14” describes an observed comparison between two reported averages. “Opening early caused an extra fourteen visitors” converts an association into a causal finding the source cannot support.
3. The ladder of certainty is a practical guide, not a law
Verbs such as claims, suggests, indicates, shows, establishes and proves often convey different attitudes to evidence. “Claims” foregrounds what someone asserts without guaranteeing its truth. “Suggests” presents a plausible interpretation. “Indicates” commonly points to evidence leaning in a direction. “Shows” often communicates direct support for a specified observation. “Proves” is much stronger and needs exceptional care.
The precise meaning depends on context. “The thermometer shows 26°C” reports a displayed reading, not a philosophical proof that every part of a room is exactly that temperature. “Her smile showed relief” is an interpretation in a narrative; body language does not supply laboratory certainty. Do not memorise a single rigid ranking and apply it without thinking.
The useful question is what exactly is supported? A source may prove a narrow mathematical proposition under stated assumptions, show a measured number, indicate a pattern and merely suggest an explanation for that pattern. Strong writers select a verb only after identifying the claim’s scope.
4. “Claims” does not necessarily mean “lies”
When a leaflet claims a new timetable will save time, a reader should recognise that it is the leaflet’s assertion. The word claims can sound sceptical in some contexts, but it is not automatically an accusation of deception. A source may make a fully accurate claim; the issue is whether the student has checked or been given grounds to accept it.
Suppose a passage reports: “The organiser claims that every child can finish the course in four weeks.” A careful answer might say “The organiser claims the course can be completed within four weeks, but the passage supplies no completion records.” This separates attribution and evidence without declaring the organiser dishonest.
If the question asks what the organiser believes, the organiser’s statement may be sufficient. If it asks whether the programme truly works for every child, a statement alone is inadequate. This difference between the content of a claim and its truth helps prevent false certainty in comprehension responses.
5. “Suggests” makes room for alternatives
“The empty seats suggest that fewer people attended” is a reasonable inference only if the seats normally represent available capacity and the observation occurred during the event. Perhaps the audience moved elsewhere. Perhaps the room was opened early. The word suggests allows the reader to consider that the clue points towards an explanation without ruling out all others.
In a narrative, a character who refuses to meet another’s eyes may seem ashamed. Yet the same behaviour might show fear, resentment or distraction. If the passage supplies an additional clue—such as the character apologising after causing damage—the shame inference strengthens. Writers who use suggests should still supply textual evidence; uncertainty is not a licence for guessing.
A good answer reads: “Her avoidance of eye contact, followed by an apology, suggests that she feels ashamed of her action.” It identifies both the behavioural clue and the interpretation. A weak answer states “She is ashamed” without showing why; another weak answer lists five possible emotions without choosing the one best supported.
6. “Indicates” often points to a patterned signal
Suppose a three-day attendance log reads 42, 47 and 51 pupils. The increase indicates an upward pattern over those days. It does not establish a permanent trend for every future week. The verb points to what the limited record can support: three successive values and their direction.
In a science report, a set of consistently lower readings in an experimental group might indicate an effect worth investigating. Its strength depends on measurement reliability, comparison conditions and sample size. In an advertisement, however, “studies indicate” may be too vague unless the study, result and limitation are identifiable.
When using indicates, complete the sentence with a precise object. “The data indicate that the measured response rose across the three observations” is clearer than “The data indicate improvement”. The second hides the definition of improvement and could refer to many different outcomes.
7. “Shows” needs an object that the source can actually show
A photograph shows a cyclist wearing a helmet if the helmet is visible. It does not show that the cyclist has never had an accident. A bar chart shows values assigned to its categories; it does not automatically show why one category is larger. An eyewitness statement shows what the witness says happened, although the truth of that memory may require corroboration.
Consider “The diagram shows that the plant grew 3 cm”. If the diagram gives two height readings and the scale is accurate, that may be a fair numerical description. “The diagram shows that the fertiliser made the plant grow 3 cm” adds a causal claim that the diagram alone may not establish. The subject is no longer the measured change but its cause.
Students can test the verb by adding “directly” in their mind: does the source directly display or state this particular fact? If not, the answer may need “suggests”, a qualification or additional evidence. This mental check is useful but not infallible: sometimes a conclusion follows through several valid reasoning steps rather than appearing explicitly on the page.
8. “Proves” is strongest when the reasoning closes the alternatives
A mathematical demonstration can prove that the sum of two particular even integers is even, or more generally that any two even integers sum to an even integer, when the argument is valid. In everyday empirical contexts, a result may strongly support a conclusion without ruling out every possible source of error or alternative explanation.
Suppose four volunteers all preferred a particular timetable. Their unanimous preference does not prove the entire school’s preference. It shows the answers of four volunteers. The sample may be biased or simply too small. A scientifically careful writer usually avoids “proves” where the design cannot justify such a sweeping statement.
Do not replace every instance of “proves” mechanically. In a fictional detective story, a narrator might say a signature proves that the letter was written by the suspect; the reader should examine whether the narrative provides genuine verification. In an argument, the word itself may reveal the writer’s confidence or rhetorical pressure even if the evidence is weaker.
9. The noun after the verb carries the real risk
“The results show a difference” and “The results show the new method is superior for all learners” use the same verb but make claims of radically different breadth. The verb cannot fix a swollen conclusion. Ask how many people, events, places and conditions the statement covers.
Numbers help reveal overreach. If a class of 30 pupils provides data, do not silently write about all Singapore students. If a test lasts one afternoon, do not claim a full-year effect. If a study measures time taken, do not claim it measures confidence or enjoyment without separate evidence.
Precision therefore has two dimensions: the degree of certainty and the extent of the claim. A cautious-sounding sentence can still be misleading if its scope is too broad. “The results may suggest that every child benefits” needs more than adding “may” when the sample is small and selective.
10. Separate observation, inference and evaluation
An observation is what the source directly reports: “The board displayed three notices.” An inference draws a plausible meaning: “The repeated notices suggest that the organisers wanted the rule noticed.” An evaluation judges adequacy or impact: “Repeating the notices may improve visibility, but without audience feedback we cannot know whether the reminders were understood.”
These levels sometimes overlap in a thoughtful paragraph, but a comprehension answer should respond to the level the question asks for. If the question says “State”, a direct fact may be enough. If it asks “What does this suggest?”, merely copying the observation may fail to explain the implication. If it asks “How effective?”, a judgement requires evidence and a reason.
A practical exercise is to colour-code your reasoning in a notebook without using decorative language in the final answer. One line states the observed detail; a second explains its likely meaning; a third states the limit. You will learn to build an answer with deliberate logical steps.
11. Direct quotation can be precise but still mislead
A quotation reproduced accurately may be torn away from its limiting clause. Imagine the source says: “Most respondents preferred the later session, although the survey reached only twelve volunteers.” Quoting only “Most respondents preferred the later session” loses a condition essential to interpreting the result.
Comprehension questions may reward a relevant reference from the text, but a fragment without context can misrepresent the author’s view. Write “Most of the twelve volunteers preferred the later session” if the number matters. If the original says “may”, do not paraphrase it as “will”. If it says “several”, do not replace that with “most”.
This is a bridge from Volume 0161 on reported speech. That chapter follows speaker and time. The present chapter follows how much a statement, observation or comparison can justify.
12. Visual-text persuasion relies on selective prominence
Imagine a poster with “90% SUCCESS!” in large type and a note underneath: “Nine of ten participants completed a five-minute puzzle in one session.” The headline is mathematically connected to the note, but its context matters. Success in a short puzzle is not evidence of long-term academic success. The large type may encourage an unwary reader to supply a broader meaning.
When explaining language for effect, name the mechanism: the prominent percentage may draw attention and create a strong impression of effectiveness; the small sample and narrow task limit the inference. Avoid an unsupported psychological claim such as “all readers will believe it”. You can describe the likely intention or effect without predicting every person’s reaction.
A useful answer is: “The large percentage makes the scheme seem highly successful, while the footnote reveals that the result comes from only ten participants in one short task.” The contrast between headline and qualification is the point, not whether the percentage is arithmetically correct.
13. Narrative certainty works differently from numerical certainty
In narrative comprehension, a character’s motives are often inferred from action, dialogue and description. Consider “Jia paused before returning the borrowed notebook. Her fingers tightened around its cover.” The passage directly depicts hesitation and a physical gesture. It may suggest reluctance, anxiety or attachment, depending on surrounding details.
If the preceding paragraph says Jia had discovered a private letter inside the book, her pause could suggest uncertainty about whether to reveal it. If the preceding paragraph says it was a gift from a friend who had moved away, the same gesture might suggest sadness. Context moves an inference from plausible to strong.
When asked why a character acted, state the inference the text supports and, where needed, refer to the particular clue. Do not invent a complete backstory or diagnose the character. The aim is to answer the writer’s created situation with sensitivity to the supplied evidence.
14. Correct a certainty shift during summary writing
An 80-word summary cannot include every detail, but it must preserve the meaning of selected points. If the source says “The pilot programme appeared to reduce waiting times in two classes,” a summary should not state “The programme eliminated waiting times throughout the school.” This changes both certainty and scope.
Compression should remove unnecessary wording, not important conditions. “A two-class pilot may have shortened queues” preserves the provisional nature and the tested group. In an 80-word task, not every qualification belongs in the summary, but those that determine the accuracy of a chosen point cannot be casually discarded.
See Volume 0148 on distinct points in summary writing. That earlier guide concerns point selection and compression; this one concerns evidence-strength wording inside each selected point.
15. In an argument, clear position does not require overclaiming
A strong discursive writer can take a firm position while conceding a limitation. Consider: “Schools should offer quiet study rooms because they provide a useful option for pupils who lack suitable space at home. However, they will not solve every obstacle to study.” The view is clear; the scope is honest.
Compare a weak absolute claim: “Every student will improve because quiet rooms always increase results.” Unless the task provides extraordinary evidence, such certainty invites an easy objection. A measured claim is often more persuasive because the writer can defend it against realistic alternatives.
Evidence calibration is not the same as constant hedging. Too many words such as “perhaps”, “maybe” and “possibly” can make a paragraph evasive. Use a clear verb and state a concrete reason. Reserve qualification for the part of the argument that actually needs it.
16. The verbs are not interchangeable with academic-sounding phrases
Students sometimes replace “says” with “postulates”, “proclaims” or “demonstrates” simply to sound advanced. A difficult verb is not automatically precise. “The author states” may be the best attribution when the text contains an explicit declaration. “The author implies” should be reserved for meaning not stated directly.
Likewise, “corroborates” means that evidence provides support or confirmation alongside other evidence; it is not simply a fancy version of “reports”. “Refutes” means showing a claim to be wrong, not merely disagreeing with it. Use these words only if the relationships are really present.
The most readable examination responses often use ordinary verbs confidently: “The table shows”, “the writer suggests”, “the speaker admits”, “the results do not establish”. They make the logical relation visible without forcing the examiner to decode unnecessary ornament.
17. A small evidence audit changes the answer
Use the fictional report: “Five of six pupils tested the route and reached class more quickly, but the group was chosen from one block.” The narrow observation is that five tested pupils were quicker on that occasion. A fair inference is that the route may be promising for that group under similar conditions. A general conclusion about every pupil in the school is not supported.
Now change one condition: suppose pupils were randomly selected from several year levels, and repeated trials showed the same direction of change. Evidence for a broader claim would improve. It would still not necessarily establish every future day, because weather, congestion or timetables could differ.
This “change one condition” practice helps students learn why a stronger verb sometimes becomes defensible. It also prevents a harmful habit: changing only the wording while leaving the weakness of the evidence untouched.
18. A model comprehension response with a confidence boundary
Original question: “What does the difference between the two attendance periods suggest about the early-opening policy?” Model response: “The higher recorded average suggests that earlier opening may have made the library easier for some pupils to use. However, the reading challenge occurred during the trial, so the figures alone do not show that the change in opening hours caused the increase.”
The response answers the inferential question, names the relevant figures and identifies a plausible alternative explanation. It does not waste space defining every statistical term. The central skill is to place a boundary around the conclusion exactly where the source’s information runs out.
Do not automatically add a caveat to every short answer if the question asks only for a stated fact. The correct amount of evaluation is determined by the command word and the source. A one-sentence literal answer may be stronger than a long paragraph that drifts beyond the task.
19. A model visual-text explanation
Fictional poster: A huge heading reads “THE FASTEST WAY TO LEARN!”, while a smaller caption says “Average task time fell in one class during a supervised practice session.” Question: “How does the headline influence the reader’s impression?”
Model response: “The absolute phrase ‘the fastest’ presents the method as superior to every alternative and may encourage readers to trust it immediately. Yet the caption describes only a change in one class’s average task time; it does not compare the method with every other approach.”
The answer isolates a linguistic device, explains the impression it may create and connects the limitation to the supplied information. It does not accuse the poster creator of fraud or assume all readers react identically. In examination reading, fairness is part of precision.
20. Compare two defensible answers, not just right and wrong
A learner might answer the attendance question with “The increase indicates a positive response to earlier opening.” Another might write, “The increase suggests greater interest in early library access, although the reading challenge may also explain the pattern.” Both recognise the direction of the data; the second more clearly protects the causal uncertainty.
Sometimes a short answer is all that the mark allocation and wording require. The student should not write a dissertation simply to demonstrate scepticism. The principle is to provide enough information to answer the question accurately. If the question asks for a reason, name one. If it asks for an evaluation, compare the evidence with the claim.
During correction, identify which part of the first answer is too broad. “Positive response” may imply a preference the attendance counts did not directly measure. Replacing it with “higher recorded attendance” describes the evidence more faithfully. A precise noun can matter as much as a precise verb.
21. The six-step source-to-sentence method
First, read the command word. Second, identify the exact supporting detail in the source. Third, separate direct observation from interpretation. Fourth, decide the strength of the inference and choose a corresponding verb. Fifth, check scope: which group, period, condition or outcome? Sixth, write the answer in straightforward language and read it once for invented certainty.
Apply the method lightly. It should become a quick mental routine rather than six long notes beside every question. At the beginning of revision, writing out the steps reveals where errors occur. Later, the same checks can happen in a few seconds.
For complex passages, keep a note of who is making each claim. An author may quote a campaigner, a researcher and a critic. A student who merges their statements into one “the author says” may misattribute the view. Correct attribution and calibrated certainty belong together.
22. Thirty original practice questions with model reasoning
The following mini-questions are not official examination extracts. For each, decide what the source says and what stronger claim is not justified. The model answers illustrate appropriate wording rather than a unique set of marking-scheme phrases.
1. Direct observation
Question. A table records 10, 12 and 14 visitors on successive days. What does it show?
Model. Visitor counts increased across those three recorded days. Reason: It does not establish a month-long trend.
2. Source attribution
Question. A supplier says its bags last forever. Which verb attributes rather than verifies?
Model. The supplier claims that the bags last forever. Reason: The statement’s truth has not been checked.
3. Small sample
Question. Three volunteers liked a new snack. Does this prove the whole school likes it?
Model. No; it shows the reported preferences of three volunteers. Reason: The sample is too narrow for that generalisation.
4. Narrative inference
Question. An anxious character repeatedly checks the door. What may the action suggest?
Model. It may suggest worry about an expected arrival. Reason: Surrounding context could establish a different motive.
5. Read a qualifier
Question. A report says “may reduce delays”. Can you rewrite it as “will eliminate delays”?
Model. No. Reason: It changes possibility into certainty and reduction into elimination.
6. Precisely reported value
Question. A screen displays 28°C. What does the screen show?
Model. A displayed temperature reading of 28°C. Reason: It need not establish identical temperature everywhere.
7. Change the baseline
Question. Two weeks differ, but one has a festival. Why avoid a causal assertion?
Model. The festival is an alternative explanation. Reason: The time periods are not necessarily comparable.
8. Authorial emphasis
Question. Why might a poster use the word “guaranteed”?
Model. To create confidence or certainty in the reader. Reason: That word is itself a persuasive claim, not evidence.
9. Critical but fair
Question. A critic disagrees with a claim. Has the critic refuted it?
Model. Not necessarily. Reason: Refutation requires a successful reason or counterexample.
10. Graphic scope
Question. One bar is higher for Class A. Can you conclude Class A worked harder?
Model. No. Reason: The graphic may show performance, not effort.
11. Observation vs explanation
Question. A pupil’s score rose from 60 to 72. State the observation.
Model. The score increased by 12 marks. Reason: The scores alone do not identify the cause.
12. Paraphrase accurately
Question. Rewrite “several respondents” without increasing certainty.
Model. A number of respondents. Reason: Do not silently turn it into “most respondents”.
13. Question type
Question. What does “What does this suggest?” request?
Model. An evidence-supported inference. Reason: Copying a bare fact may be insufficient.
14. Tighten a claim
Question. “The trial proves every student benefits” becomes what?
Model. “The trial suggests a possible benefit for the tested students.” Reason: Its scope and certainty now match a limited trial.
15. Limiting clause
Question. Why retain “among twelve volunteers” in a summary?
Model. It identifies the limited group. Reason: Omitting it may imply a wider population.
16. Evidence hierarchy
Question. Is an organiser’s own statement a measurement?
Model. No. Reason: It reports the organiser’s assertion.
17. Contradiction
Question. A headline says “all pupils”; footnote says “nine tested pupils”. What is wrong?
Model. The headline extends beyond the footnote’s population. Reason: The scope changes without support.
18. Reading behaviour
Question. A character smiles after a difficult conversation. What can be inferred cautiously?
Model. The smile may indicate relief or reassurance. Reason: Use adjacent context to determine which is supported.
19. Neutral reporting
Question. Use a verb to report that an article expressly writes “the library closes at six”.
Model. The article states that the library closes at six. Reason: No inference about hidden intentions is needed.
20. Reject false proof
Question. Does four identical observations prove the instrument has no systematic bias?
Model. No. Reason: Repeatability alone cannot rule out a common error.
21. Relevant contrast
Question. A method is faster but causes more mistakes. What must an evaluation include?
Model. Both speed and accuracy. Reason: A one-sided claim of improvement may be misleading.
22. Citation logic
Question. Does quoting a politician establish that the politician’s promise was fulfilled?
Model. No. Reason: The quotation establishes what was promised.
23. Strong reason
Question. Why is “may suggest” not a replacement for evidence?
Model. A qualifier does not supply supporting observations. Reason: An inference still needs a textual basis.
24. Mathematical proof
Question. When may “proves” be appropriate?
Model. When a logically valid argument establishes the stated proposition. Reason: Proof applies to the exact claim derived.
25. Uncertain memory
Question. An eyewitness recalls a number but says “approximately”. How report it?
Model. The witness estimated the number. Reason: Do not replace the estimate with an exact count.
26. Speaker distinction
Question. A report quotes a critic’s view. Who holds the view?
Model. The critic, unless the author explicitly adopts it. Reason: Attribution protects voice ownership.
27. Counterevidence
Question. A new dataset has one case that breaks an “always” claim. What happens?
Model. The universal claim is false in that stated setting. Reason: One valid counterexample can refute an unrestricted universal proposition.
28. Good summarising
Question. What should remain when compressing a conditional claim?
Model. The relevant condition and the supported outcome. Reason: Compression should preserve the truth conditions.
29. Rhetorical promise
Question. Why is “completely safe” stronger than “tested under these conditions”?
Model. It claims a universal outcome rather than a limited test result. Reason: The second states the actual scope.
30. Final calibration
Question. A pilot recorded higher attendance, but a campaign ran simultaneously. Give one balanced sentence.
Model. Attendance was higher during the pilot, although the figures do not isolate the effect of opening hours. Reason: The sentence states the observation and its causal limit.
23. A full revision passage: separate three speakers
Original teaching passage: “The community centre invited twelve volunteers to try a quiet study room. Nine reported that they could focus better during the first session. The organiser called the result ‘conclusive proof that quiet rooms improve every student’s performance’. A volunteer said the quieter setting helped her complete a difficult worksheet. The centre’s evaluator noted that the participants were not randomly selected and that test scores were not collected.”
Identify three different voices. The organiser makes a sweeping effectiveness claim. The volunteer reports her experience, which is meaningful but personal. The evaluator describes design limits and absent outcome data. These statements may coexist without one being a lie: they serve different purposes and rely on different kinds of evidence.
A high-quality response might say: “Most of the volunteers reported better focus, but the study involved a small self-selected group and did not measure academic performance. The results therefore suggest that the room may help some users concentrate; they do not prove an improvement for all students.” This version respects the data and draws a plausible limited inference.
24. Four questions on that passage
Question A: What numerical result did the centre obtain? Answer: Nine of twelve volunteers reported better focus in the first session. This is a reported perception, not a measured exam-grade increase.
Question B: Why is “conclusive proof” too strong? Answer: The group was small and not randomly selected, and the centre did not collect performance scores. The claim about every student’s performance goes beyond the outcome measured.
Question C: How does the volunteer’s comment contribute? Answer: It supplies one concrete personal experience consistent with the idea that the room may be helpful, but it does not establish how all students would respond.
Question D: Rewrite the organiser’s claim responsibly. Answer: “The trial suggests the quiet room may support concentration for some volunteers, and further study would be needed to examine academic performance.” The revision reports a possible benefit without inventing an effect size.
25. A paragraph that moves from claim to judgement
Begin with a clear source statement: “The advertisement claims that the service saves time.” Then state the supplied evidence: “Its example shows that one user finished a particular task five minutes faster.” Next identify the limitation: “The text does not provide measurements for other users or tasks.” Finally reach a fair judgement: “The evidence suggests a possible benefit but cannot justify the universal promise.”
These four moves can form a short evaluation, but their order is flexible. If the question asks for one reason, you may need only the relevant limitation. If it asks you to evaluate the advertisement, the fuller structure helps. The important point is that each sentence performs a distinct job.
Do not turn the structure into a stock paragraph with empty nouns. Replace “the evidence”, “the benefit” and “the claim” with the actual terms of the task. Specificity is how a reader knows you understood this text instead of any text.
26. Two likely wrong paths on exam day
One wrong path is to treat every text as a conspiracy. A poster’s persuasive headline is not necessarily false. A writer may intend to encourage readers using a memorable expression. Your task is to explain how the choice works and whether the available evidence supports the meaning the question asks about.
The opposite error is to accept every emphatic phrase at face value. A confident narrator, researcher or organiser may be limited, biased, mistaken or deliberately selective within the passage. Read what the source actually supplies. An examiner is assessing your understanding of text and evidence, not your willingness to agree with a powerful voice.
Good readers keep the two paths apart. They neither accuse without evidence nor surrender judgement to confidence. The useful middle is an explanation precise enough to withstand a simple follow-up: “Which detail supports that, and which detail limits it?”
27. How to practise the skill for seven days
Day one: collect ten verbs from ordinary school notices and identify who is speaking. Day two: rewrite five absolute claims with accurate limits using the original source facts. Day three: distinguish observation, inference and evaluation in a short narrative. Day four: compare the headline and small print of a fictional poster and explain their different effects.
Day five: practise compressing a qualified paragraph without dropping its essential conditions. Day six: give one concise inferential answer and one evaluative answer to the same passage. Day seven: attempt a mixed K300 comprehension practice from an authorised school resource, then classify your errors as mistaken evidence, mistaken scope, mistaken speaker or excessive certainty.
This routine is a suggested workshop, not a prescribed SEAB activity. Its purpose is to make a specific mental skill reliable, so students can bring it into unfamiliar passages rather than remembering only the examples from this page.
28. Questions parents and learners ask
Should I avoid “proves” in every answer? No. Use it when the reasoning warrants the particular claim, including valid mathematical demonstrations. In empirical or narrative inference, be cautious about claims of absolute certainty.
Is “suggests” always weaker than “indicates”? The difference depends on sentence and context. Both can convey an evidence-based inference. What matters is that the claim’s strength and scope are fair.
Should every answer mention limitations? No. Give what the question asks. A literal factual question may need a direct answer; an evaluative question may need conditions or alternatives.
Can a paraphrase change the grammatical structure? Yes, provided the meaning, agency, time, numbers and certainty remain intact.
Does this chapter replace K300 practice papers? No. It is a focused reading and writing skill within the broader official syllabus.
29. Continue the learner route
Use the G3 SEC Learner Guide for examination planning, the PSLE Learning Guide for the foundational move from literal meaning to evidence-based inference, and the Secondary English Learning Hub for wider reading, grammar, writing and oral work.
For previous practice, Volume 0161 focuses on reported speech; Volume 0181 teaches sentence connections; and Volume 0184 demonstrates why real-world evidence requires disciplined interpretation.
The Mathematics Hub and Science Hub develop parallel habits of checking quantities and evidence. Consult the official 2027 K300 syllabus for examination requirements rather than treating any educational example as a forecast.
30. Finish with a sentence the evidence can carry
Strong comprehension is not a contest to find the most impressive verb. It is an exercise in giving an answer that is no broader, narrower, weaker or stronger than the relevant facts allow. The reader should be able to ask where the idea came from and receive a precise reply.
Start with what the text directly states or depicts. Make the necessary inference. If the conclusion is conditional, keep that condition in view. If the writer uses a forceful promise, explain its effect without assuming the promise is true. If a mathematical result is proven, say exactly what has been established.
When students learn to make this distinction, their answers become shorter, clearer and more defensible. That is a lasting English skill: use words not only to express an idea, but to show how well the idea is supported.
