Small Group Tutorials

Here to help students catch up, keep up, and move ahead. Book a consultation here.

How to Perform in the new G3 SEC Examinations | Learner’s Guide Vol 1450 | EMS Refill-Station Error-Detection Masterclass — Chemistry, Cost Models and Public Claims

A final recommendation can sound clear while resting on a chain of errors. A technician confuses filtration with purification. An accountant compares a first-month price against a full-year price. A writer changes ‘some respondents’ into ‘all students’. Every sentence seems small, yet the combination produces a confident conclusion that nobody has properly established. This learner’s guide teaches you to catch the earliest wrong move, not merely disagree with the ending.

You will work through a fictional school refill-station proposal from three directions: Science asks what the sample and tests demonstrate; Mathematics asks whether the numbers have matching units, time horizons and assumptions; English asks whether the final language accurately represents the evidence. The method is designed for advanced G3 examination practice while remaining understandable to a student who needs to rebuild foundational habits first. No example is a statement about a real Singapore school, supplier, installation or drinking-water result.

Quick summary: State the exact claim; identify the subject-specific evidence needed; separate observed conditions from an imagined general rule; match all units and cost periods; recalculate with the same assumptions; check sample sizes and denominators; qualify unsupported language; write a short verdict that identifies what is known, not known and needed next.

1. How this article fits the SEC and earlier learner guides

The Singapore-Cambridge Secondary Education Certificate begins in 2027. It replaces the previous N- and O-Level framework, with students taking eligible subjects at their respective G1, G2 or G3 levels. For 2027 standalone G3 subjects, SEAB lists English Language as K300 (formerly 1184), Mathematics as K310 (formerly 4052), and Chemistry as K324 (formerly 6092). These are distinct examined subjects; EMS here is a study method, not the name of a single SEC examination paper.

The official formats remain different. K300 Paper 2 Comprehension is 50 marks and 35% of English; K310 Mathematics has two 90-mark papers, each 2 hours 15 minutes and each worth 50%; K324 Chemistry has a 40-mark MCQ paper worth 30%, an 80-mark structured/free-response paper worth 50%, and a 40-mark practical worth 20%. An integrated home exercise does not replace timed practice for each actual subject. Check the official 2027 SEAB G3 index for the current papers.

For full subject teaching, return to Volume 1447 on Chemistry separation and fair trials, Volume 1448 on linear cost models, and Volume 1449 on comprehension and attribution. This volume assumes those skills and asks whether you can detect a mixed error before it misleads a decision.

2. The invented decision file

Imagine that a school committee has received two equipment proposals and a training-laboratory report. Plan A charges a setup fee of $180, $28 per month and $0.06 for each refill. Plan B charges a setup fee of $360, $22 per month and $0.04 for each refill. We assume the same demand, identical counting method, no discounts, no taxes and no other relevant charges. These figures are fictional teaching data, not supplier quotations.

The pilot bulletin records average daily counter totals of 72 refills in a first monitored week and 105 in the fourth. A separate voluntary consultation records 156 supportive submissions out of 240. A technician records three unplanned interruptions lasting 8, 12 and 9 minutes and a planned six-minute calibration. A school laboratory also separates a visibly cloudy model mixture for teaching, using filtration and then a demonstration of distillation on another prepared sample. None of these classroom operations certifies a real drinking-water supply.

A draft recommendation says: ‘The filter proves the water is pure; the school already has 105 separate daily users; 65% of all students demand expansion; Plan B is always cheaper; and the trial had 35 minutes of equipment failures. We should announce that a new station will definitely be built.’ The document is useful as an examination exercise precisely because nearly every clause has a different kind of fault.

3. Start with the claim, not the mood

Before repairing a statement, classify it. Some are measurements, some are calculations, some are inferences, and some are proposed actions. A candidate who reacts only to the positive or negative tone risks overlooking the mechanism of error. ‘Always cheaper’ is an absolute claim about cost over all periods and usages. It cannot be proved by showing a single year’s result. ‘Pure’ is a scientific property that cannot be established by visual clarity alone. ‘Definitely’ is a decision word requiring authority and an actual decision.

A good correction avoids the opposite overstatement. If the science evidence does not prove purity, it does not necessarily prove contamination. If 156 respondents support examining expansion, that does not prove the whole school opposes it. If Plan A is cheaper initially while Plan B can be cheaper later, calling either one ‘bad’ ignores the conditions. Better answers draw a boundary around what the data justify.

4. The mathematical backbone: compare like with like

Define n as the number of completed refills per month, assumed constant for twelve months. For Plan A, annual cost A(n)=180+12(28)+12(0.06n)=516+0.72n. For Plan B, annual cost B(n)=360+12(22)+12(0.04n)=624+0.48n. Equal costs require 516+0.72n=624+0.48n, hence 0.24n=108 and n=450 monthly refills. Under these assumptions, A is cheaper below 450, the plans tie at 450, and B is cheaper above 450.

But first-month comparison has different intercepts. A₁(n)=208+0.06n and B₁(n)=382+0.04n. Equality becomes 0.02n=174, or n=8700 refills in the first month. For 2,000 refills in that first month, A costs $328 and B costs $462. For twelve months at 2,000 refills each month, A costs $1,956 and B costs $1,584. These answers do not contradict each other: they answer questions about different time periods.

Capacity is separate. The lower annual cost of B at a chosen demand is not enough if the device cannot physically process that demand in the available opening hours. Equally, a supposed savings calculation should not assume that every refill replaces one disposable bottle without evidence. Algebra can model assumptions precisely; it cannot make those assumptions true.

5. The scientific backbone: what the apparatus establishes

Filtration separates an insoluble solid from a liquid when suitable filter media retain the particles. A clear filtrate may still contain dissolved substances or microorganisms. Distillation uses differences in boiling point or volatility to separate substances, and the condensed liquid needs appropriate collection and handling; a classroom demonstration is not a certified drinking-water test. Chromatography can help separate components and compare movement under specified conditions, but visual patterns alone should not be treated as a comprehensive safety verdict.

Chemistry practical answers also require controls. If two methods are compared, identify an independent variable, a dependent measure and relevant constants. Repeat where appropriate. Report what was actually observed, not what would be desirable for the project. The 2027 K324 Chemistry practical paper assesses investigation and data skills; a vague line such as ‘make it fair’ is not a complete experimental method.

6. The language backbone: evidence must survive editing

English Paper 2 trains candidates to locate evidence, infer carefully and explain language effects. The same discipline is essential when revising a report. ‘The counter recorded 105 refills on an average monitored day in week four’ is not interchangeable with ‘105 pupils used the machine daily’. One sentence names events and an averaging period. The other asserts distinct people and a universal daily pattern.

Likewise, ‘156 of 240 voluntary respondents supported examining another location’ is not equivalent to ‘65% of the school demands installation’. The first denominator is 240 submitted responses; the second would require a different data-collection frame. Precise public writing is a form of reasoning. Avoid inventing motives for a writer just because a number was presented enthusiastically.

7. Thirty-two error-detection decisions

Audit 01 — A transparent filtrate must be pure.

Corrected finding: Transparency concerns visible suspended matter. It does not demonstrate the absence of dissolved species or biological hazards. Examination reasoning: The examiner is testing whether appearance is being confused with composition. Try it independently: Specify a relevant additional test and the particular property it would examine. Before moving to the next item, mark whether the original fault was in observation, definition, unit, sample, time period, causal inference or decision authority. That diagnosis will guide a more useful correction than simply memorising the model wording.

Audit 02 — A filter removes all salts dissolved in water.

Corrected finding: Ordinary filtration of a solution does not separate dissolved salt ions as if they were insoluble solids. Examination reasoning: Use the distinction between dissolved particles and an insoluble residue. Try it independently: Explain why a second separation method may be needed. Before moving to the next item, mark whether the original fault was in observation, definition, unit, sample, time period, causal inference or decision authority. That diagnosis will guide a more useful correction than simply memorising the model wording.

Audit 03 — Distillation automatically produces safe drinking water.

Corrected finding: Distillation separates based on physical properties, but neither an invented school experiment nor a simple description certifies potable water. Examination reasoning: A separation principle is not a full safety assurance protocol. Try it independently: Replace ‘safe’ with the specific observation actually made. Before moving to the next item, mark whether the original fault was in observation, definition, unit, sample, time period, causal inference or decision authority. That diagnosis will guide a more useful correction than simply memorising the model wording.

Audit 04 — Any two chromatograms can be compared directly.

Corrected finding: Comparison needs compatible conditions such as solvent, stationary phase and reference behaviour; distances alone are not enough. Examination reasoning: The relevant variables matter to the conclusion. Try it independently: State what should remain controlled before using Rf values. Before moving to the next item, mark whether the original fault was in observation, definition, unit, sample, time period, causal inference or decision authority. That diagnosis will guide a more useful correction than simply memorising the model wording.

Audit 05 — A higher Rf value means a substance is more toxic.

Corrected finding: Rf concerns relative migration in a chromatographic system, not toxicity. Examination reasoning: The physical measurement does not carry a health judgement. Try it independently: Separate identification-related evidence from hazard assessment. Before moving to the next item, mark whether the original fault was in observation, definition, unit, sample, time period, causal inference or decision authority. That diagnosis will guide a more useful correction than simply memorising the model wording.

Audit 06 — The clearest sample is guaranteed to be the cleanest.

Corrected finding: Visible clarity is only one property and ‘cleanest’ is undefined without a measure. Examination reasoning: An imprecise adjective hides the missing operational definition. Try it independently: Describe a measurable criterion instead of an impression. Before moving to the next item, mark whether the original fault was in observation, definition, unit, sample, time period, causal inference or decision authority. That diagnosis will guide a more useful correction than simply memorising the model wording.

Audit 07 — One trial proves a method will always work.

Corrected finding: An individual successful observation has limited scope and may not cover changes in sample composition or conditions. Examination reasoning: Replication and the range of tested conditions affect generalisation. Try it independently: Propose a sensible repeat and an altered-condition check. Before moving to the next item, mark whether the original fault was in observation, definition, unit, sample, time period, causal inference or decision authority. That diagnosis will guide a more useful correction than simply memorising the model wording.

Audit 08 — Changing two variables makes an experiment more realistic.

Corrected finding: Simultaneous changes can make cause-and-effect attribution difficult. Examination reasoning: Realism is not a substitute for interpretability in a controlled comparison. Try it independently: Identify the variable you would deliberately change first. Before moving to the next item, mark whether the original fault was in observation, definition, unit, sample, time period, causal inference or decision authority. That diagnosis will guide a more useful correction than simply memorising the model wording.

Audit 09 — The pilot had four faults.

Corrected finding: It recorded three unplanned interruptions and one planned calibration check, not four recorded failures. Examination reasoning: Similar-looking time entries have different categories. Try it independently: Write a table that separates planned work from unexpected downtime. Before moving to the next item, mark whether the original fault was in observation, definition, unit, sample, time period, causal inference or decision authority. That diagnosis will guide a more useful correction than simply memorising the model wording.

Audit 10 — The faults lasted 35 minutes.

Corrected finding: The three unplanned interruptions totalled 29 minutes. Adding six minutes of scheduled calibration changes the category. Examination reasoning: A correct addition can still be the wrong answer if its terms do not belong together. Try it independently: Explain which measurement a 35-minute total would actually represent. Before moving to the next item, mark whether the original fault was in observation, definition, unit, sample, time period, causal inference or decision authority. That diagnosis will guide a more useful correction than simply memorising the model wording.

Audit 11 — The refill counter measured litres dispensed.

Corrected finding: A count of refill events is not a measured volume. Examination reasoning: Units are the first defence against a false inference. Try it independently: State what extra sensor or record would be needed. Before moving to the next item, mark whether the original fault was in observation, definition, unit, sample, time period, causal inference or decision authority. That diagnosis will guide a more useful correction than simply memorising the model wording.

Audit 12 — The fourth week had 105 unique users daily.

Corrected finding: The figure describes average refills, not identifiable individuals. Examination reasoning: Counts of actions and counts of people are not identical. Try it independently: Paraphrase with the correct observational noun. Before moving to the next item, mark whether the original fault was in observation, definition, unit, sample, time period, causal inference or decision authority. That diagnosis will guide a more useful correction than simply memorising the model wording.

Audit 13 — An increase from 72 to 105 is 105 extra refills.

Corrected finding: The absolute increase is 33 average daily counts; 105 is the later total. Examination reasoning: Beware ‘to’ versus ‘by’ when writing about change. Try it independently: Write both the beginning and end values in the same sentence. Before moving to the next item, mark whether the original fault was in observation, definition, unit, sample, time period, causal inference or decision authority. That diagnosis will guide a more useful correction than simply memorising the model wording.

Audit 14 — 65% of all pupils voted for another station.

Corrected finding: 156/240 = 65% of voluntary respondents favoured examining a location, not necessarily all pupils or actual installation. Examination reasoning: Both the sample and the action in the question were enlarged. Try it independently: Repair the denominator and the decision verb. Before moving to the next item, mark whether the original fault was in observation, definition, unit, sample, time period, causal inference or decision authority. That diagnosis will guide a more useful correction than simply memorising the model wording.

Audit 15 — Every non-supporter opposed the plan.

Corrected finding: The other 84 responses may include disagreement and non-expression of support; the detail does not establish a uniform opposition category. Examination reasoning: Do not manufacture categories because arithmetic leaves a remainder. Try it independently: Use a neutral description of the remaining responses. Before moving to the next item, mark whether the original fault was in observation, definition, unit, sample, time period, causal inference or decision authority. That diagnosis will guide a more useful correction than simply memorising the model wording.

Audit 16 — A survey of 240 responses proves the whole school agrees.

Corrected finding: The sample size alone does not remove voluntary-response bias or show population coverage. Examination reasoning: Representativeness depends on sampling method as well as count. Try it independently: Name an improved collection plan before generalising. Before moving to the next item, mark whether the original fault was in observation, definition, unit, sample, time period, causal inference or decision authority. That diagnosis will guide a more useful correction than simply memorising the model wording.

Audit 17 — Plan A costs only $28 per month.

Corrected finding: The $28 figure is one component; refill charges and an initial setup fee also matter. Examination reasoning: A partial price is not the same as total cost. Try it independently: Give the complete first-month expression. Before moving to the next item, mark whether the original fault was in observation, definition, unit, sample, time period, causal inference or decision authority. That diagnosis will guide a more useful correction than simply memorising the model wording.

Audit 18 — Plan B is always cheaper because it charges $0.04 per refill.

Corrected finding: Plan B has a lower variable rate but a larger one-time charge; the time horizon and usage determine total cost. Examination reasoning: Compare full expressions, not one coefficient. Try it independently: Explain why the initial advantage can reverse. Before moving to the next item, mark whether the original fault was in observation, definition, unit, sample, time period, causal inference or decision authority. That diagnosis will guide a more useful correction than simply memorising the model wording.

Audit 19 — First-month and annual costs can be compared directly.

Corrected finding: They cover different durations and usually different volumes of service. Examination reasoning: Equivalent time horizons are required for a fair total-cost comparison. Try it independently: State a consistent period and assumption before calculating. Before moving to the next item, mark whether the original fault was in observation, definition, unit, sample, time period, causal inference or decision authority. That diagnosis will guide a more useful correction than simply memorising the model wording.

Audit 20 — At 2,000 monthly refills, Plan A is always cheapest.

Corrected finding: Plan A is cheaper for the first month under given assumptions, while Plan B is cheaper over twelve months at sustained demand. Examination reasoning: A conditional statement must keep its condition. Try it independently: Present both numbers and the time horizons. Before moving to the next item, mark whether the original fault was in observation, definition, unit, sample, time period, causal inference or decision authority. That diagnosis will guide a more useful correction than simply memorising the model wording.

Audit 21 — For one year, break-even is 8,700 monthly refills.

Corrected finding: The 8,700 figure is the first-month break-even; annual break-even at constant monthly use is 450. Examination reasoning: Reusing a result outside its equation is a transfer failure. Try it independently: Show the equation that gives 450. Before moving to the next item, mark whether the original fault was in observation, definition, unit, sample, time period, causal inference or decision authority. That diagnosis will guide a more useful correction than simply memorising the model wording.

Audit 22 — Any negative solution would make operational sense.

Corrected finding: A mathematical result must also satisfy domain constraints such as non-negative integer refills. Examination reasoning: Algebra may produce values that have no feasible interpretation. Try it independently: Check domain and reject an impossible quantity explicitly. Before moving to the next item, mark whether the original fault was in observation, definition, unit, sample, time period, causal inference or decision authority. That diagnosis will guide a more useful correction than simply memorising the model wording.

Audit 23 — If the costs tie at 450, they tie at every usage.

Corrected finding: The two lines intersect at one monthly usage under the stated annual model; the slopes differ. Examination reasoning: Equal at a point does not mean identical functions. Try it independently: Compare at 100 and 1,000 monthly refills. Before moving to the next item, mark whether the original fault was in observation, definition, unit, sample, time period, causal inference or decision authority. That diagnosis will guide a more useful correction than simply memorising the model wording.

Audit 24 — A cheaper contract automatically guarantees better service.

Corrected finding: Cost does not alone establish reliability, capacity, maintenance quality or safety. Examination reasoning: An optimisation criterion should not silently become a full judgement. Try it independently: List two operational measures still needed. Before moving to the next item, mark whether the original fault was in observation, definition, unit, sample, time period, causal inference or decision authority. That diagnosis will guide a more useful correction than simply memorising the model wording.

Audit 25 — One extra refill always means one plastic bottle saved.

Corrected finding: Displacement is not measured by a refill counter, and behaviour varies. Examination reasoning: A causal environmental claim needs a separate evidence chain. Try it independently: Describe a way to estimate avoided bottle purchases responsibly. Before moving to the next item, mark whether the original fault was in observation, definition, unit, sample, time period, causal inference or decision authority. That diagnosis will guide a more useful correction than simply memorising the model wording.

Audit 26 — The slogan proves the programme has succeeded.

Corrected finding: A slogan communicates an intention. It is not a measured outcome. Examination reasoning: Promotional language and empirical evidence serve different roles. Try it independently: Provide a neutral replacement based on recorded facts. Before moving to the next item, mark whether the original fault was in observation, definition, unit, sample, time period, causal inference or decision authority. That diagnosis will guide a more useful correction than simply memorising the model wording.

Audit 27 — The poster promised the machine would never fail.

Corrected finding: The description of school-day access is not an explicit reliability warranty. Examination reasoning: Unfair quotation is itself an English comprehension error. Try it independently: State the actual hours and report rule instead. Before moving to the next item, mark whether the original fault was in observation, definition, unit, sample, time period, causal inference or decision authority. That diagnosis will guide a more useful correction than simply memorising the model wording.

Audit 28 — The technician admitted that the water was unsafe.

Corrected finding: The technician’s caution about visual inspection does not establish contamination. Examination reasoning: ‘Not demonstrated safe’ differs from ‘demonstrated unsafe’. Try it independently: Write both claims and distinguish their evidence. Before moving to the next item, mark whether the original fault was in observation, definition, unit, sample, time period, causal inference or decision authority. That diagnosis will guide a more useful correction than simply memorising the model wording.

Audit 29 — The committee has already approved expansion.

Corrected finding: The pilot announcement says the trial does not decide future installation. Examination reasoning: A proposed examination and formal authorisation are distinct. Try it independently: Identify who would need to issue an actual decision. Before moving to the next item, mark whether the original fault was in observation, definition, unit, sample, time period, causal inference or decision authority. That diagnosis will guide a more useful correction than simply memorising the model wording.

Audit 30 — The report can remove all qualifiers to sound confident.

Corrected finding: Removing ‘monitored’, ‘voluntary’, ‘under these assumptions’ and ‘may’ changes the strength of conclusions. Examination reasoning: A qualifier can be the information that makes a sentence true. Try it independently: Restore four lost qualifiers in a practice paragraph. Before moving to the next item, mark whether the original fault was in observation, definition, unit, sample, time period, causal inference or decision authority. That diagnosis will guide a more useful correction than simply memorising the model wording.

Audit 31 — A conclusion must be either full approval or full rejection.

Corrected finding: A limited recommendation can identify evidence gaps, conditions and further measurements. Examination reasoning: Exam evaluation rewards the scope of the available information. Try it independently: Write a next-step decision with one reason and one condition. Before moving to the next item, mark whether the original fault was in observation, definition, unit, sample, time period, causal inference or decision authority. That diagnosis will guide a more useful correction than simply memorising the model wording.

Audit 32 — If calculations are correct, the final recommendation must be correct.

Corrected finding: Correct arithmetic is necessary but not sufficient: wrong assumptions or unsupported science can still defeat the conclusion. Examination reasoning: Cross-disciplinary error checking protects the meaning of a number. Try it independently: Identify the first unsupported premise in a fluent report. Before moving to the next item, mark whether the original fault was in observation, definition, unit, sample, time period, causal inference or decision authority. That diagnosis will guide a more useful correction than simply memorising the model wording.

8. A full worked repair of the committee’s argument

Original: ‘We know students will use the station because average daily refills increased and 65% of everyone voted for it.’ The first clause moves from measured refill events to predicted usage by individuals. The second falsely changes a voluntary sample into the full school. A defensible rewrite is: ‘Recorded refill events increased between the first and fourth monitored weeks. In a voluntary consultation, 156 of 240 respondents supported examining another location. These findings suggest some interest, but neither establishes how many distinct students would use an additional station.’

Original: ‘The filter showed pure water, and costs will definitely fall under B.’ Two separate conditions are missing. First, filtration of a model mixture illustrates separation but does not establish comprehensive purity. Second, the savings depend on an annual usage assumption. Revised: ‘The teaching experiment demonstrated removal of insoluble material under its stated conditions. Under a twelve-month model with 2,000 refills per month and the stated charges, Plan B’s modelled cost is lower than Plan A’s. Further safety, capacity and service evidence would be needed before operational approval.’

Now ask why that second revision is better. It narrows the scientific outcome, names a specific mathematical horizon, retains the usage assumption, and separates cost from authorisation. It may be longer than the original, but it is not cluttered. Every clause fixes a particular error. Once the reasoning is complete, a shorter public summary can be written without losing meaning.

9. A model decision brief suitable for disciplined practice

Purpose: Provide an evidence-limited recommendation to a fictional school committee. Observed information: Counter activity increased over the monitored period, but the counter does not identify unique users or measure waste avoided. A voluntary consultation recorded majority support among respondents for exploring another location, not binding approval for construction. The technician reported 29 minutes of unplanned interruption and a separate planned calibration.

Scientific position: The laboratory exercise demonstrated selected separation techniques using teaching samples. It does not provide a drinking-water safety certificate. Cost position: Under a constant-demand, twelve-month model, B overtakes A above 450 refills per month and is cheaper at 2,000 monthly refills. However, capacity, maintenance quality, safety compliance and actual usage remain to be investigated. Decision: Do not claim that expansion is proved. Request the missing evidence, verify the assumptions and decide through the appropriate school process.

A good recommendation does not need to sound timid. It can be decisive about what work should be completed next. The important difference is between certainty about the next evidence-gathering step and certainty about an outcome the evidence has not yet determined.

10. Three subject-specific answer styles

In Chemistry: ‘Filtration removes insoluble particles because these are retained by the filter medium, but dissolved substances can pass through.’ This answer states a mechanism and a limitation. In Mathematics: ‘For constant monthly demand n over twelve months, A(n)=516+0.72n and B(n)=624+0.48n. Their break-even value is n=450.’ This answer defines units, models, works and interprets.

In English: ‘The council favoured examining another location rather than promising construction; the phrase limits what the responses can be said to authorise.’ This answer grounds interpretation in wording. The three styles are not interchangeable, yet all require the same careful movement from given information to justified statement.

11. A timed EMS revision circuit

Start with twelve minutes of Science: distinguish separation methods, annotate one hypothetical observation table, name the controlled variables and write a qualified conclusion. Then use twelve minutes of Mathematics to set up the two cost functions, solve the annual crossover, compare first-month and annual examples, and inspect feasibility. Finish with twelve minutes of English: repair six claims in the draft report and summarise the decision limits in approximately 80 words. This 36-minute circuit is a home-study exercise, not an official SEC combined-paper format.

After completion, spend twenty minutes reviewing errors by type. Did a Chemistry answer assume visual clarity meant purity? Did a cost line omit multiplication by twelve? Did a comprehension phrase silently broaden a population? For a delayed check, repeat the same skills two days later using changed numbers and a different sample description. Transfer matters more than remembering the order of the model answers.

12. Another worked model with changed assumptions

Suppose usage is 300 refills per month for a year. Plan A costs 516+0.72(300)=$732, while Plan B costs 624+0.48(300)=$768. A is $36 cheaper. At 800 monthly refills, A costs $1,092 while B costs $1,008, so B is $84 cheaper. Both calculations follow from the same model. The numbers reverse the preferred option because demand crosses the 450-refill threshold.

Now suppose the school chooses not to include setup fees because it is comparing only continuing costs after installation. That is a new question, not the original model. The recurring annual charges would be 336+0.72n for A and 264+0.48n for B, so B is lower at every non-negative usage in this restricted comparison. It would be misleading to present that recurring-only result as the full first-year cost. Every calculation begins with a definition of what is included.

13. What the examiner means by ‘evaluate’

To evaluate is not necessarily to reject. It is to judge evidence with criteria. In the Science paper, compare whether the method tests the intended property. In Mathematics, examine whether assumptions, scale, accuracy and constraints support the numerical conclusion. In English, decide whether a writer’s claims and rhetorical choices are justified by the supplied material. Good evaluation names both strengths and limitations where relevant.

Practice this three-sentence structure: ‘This finding supports X because of Y. It cannot yet establish Z because of limitation L. A suitable next check would be M.’ The structure is useful for revision but should not be forced into every short answer. One-mark questions need concise responses, while longer explanations may require a full chain of reasoning.

14. Keeping the learner’s first weak link visible

A student who repeatedly writes an unsupported conclusion may not need another hundred questions. They may need to practise stopping at the step where observations turn into claims. Another student may understand the evidence but struggle with algebraic coefficients, or be unable to state the limitation in standard English. Different errors deserve different teaching. A high mark in a familiar worksheet is less reassuring than an accurate response to a changed task.

For students moving from PSLE, keep the entry-route terminology separate. Posting Groups organise Secondary 1 admission; G1, G2 and G3 describe subject levels. The post-PSLE learning plan and Sengkang school choices guide can help families make a gradual bridge rather than expecting a Secondary 1 student to perform at graduating G3 SEC standard immediately.

15. Family practice without turning the home into an exam hall

Ask the learner to explain one correction aloud. If the sentence was ‘the water is safe’, ask what was actually tested. If the result was ‘B is cheaper’, ask over what time period. If the statement was ‘everyone agrees’, ask who was surveyed. Let the student locate the premise and repair it independently. The conversation should reveal thinking, not reward the speed of guessing the parent’s intended answer.

The English owner addresses comprehension and the writing of accurate claims. The Mathematics owner addresses algebra, units and modelling. The Science owner is specifically for Primary 3 to Primary 6 Science; it is a useful PSLE bridge, not a claim that eduKate Sengkang currently offers secondary Science tuition. For G3 Chemistry requirements, rely on the official syllabus and the dedicated learner guide.

16. Revision questions that test transfer

How would your conclusion change if the survey included every student? What if the refill count recorded bottle IDs as well as events? What if maintenance charges were added to Plan B? What if the filtration demonstration removed a different class of material? What if the technician’s note had no dates? The answer to each begins by naming the changed assumption, then revising only the conclusions that depend on it.

Do not make every change propagate through every part of the report. A different survey sampling method could strengthen a preference inference without changing the Chemistry separation principle. A new monthly fee could change the break-even calculation without establishing that the water is safe. Integrated reasoning means knowing what is connected and what remains independent.

17. Sources and continuation

Official references: SEC explanation, 2027 G3 syllabuses, English K300, Mathematics K310, and Chemistry K324. For subject-level practice, follow Volume 1447, Volume 1448 and Volume 1449. This article is an integrated correction exercise, not a substitute for the three individual guides.