A wet patch beneath a classroom window is not an explanation. It is an observation. Water might have condensed from humid air on a cold surface; rain might have entered through a defective seal; a cleaner might have left moisture behind; or more than one process might be involved. The important skill is not to guess the most dramatic story. It is to decide what the evidence can support.
This guide helps G3 learners combine Science, Mathematics and English to investigate a cold-window report. We begin with simple observations, build a defensible explanation of phase change, calculate only what the readings justify, and finish with a careful recommendation to a school. Nothing in the fictional case study is an actual report about a Sengkang building.
By the end, you should be able to separate a measured quantity from an inferred cause, identify useful controls, interpret baseline-corrected data, recognise language that overclaims and write a calm decision that explains what is known, what is uncertain and what should be checked next.
- Science: identify water vapour, condensation, cooling, rainfall and possible infiltration without confusing them.
- Mathematics: calculate net changes, rates, area-normalised figures and fair comparisons while retaining the correct units.
- English: qualify claims, distinguish observation from interpretation and make an evidence-based recommendation.
- Exam technique: read the exact command word and answer the question asked, not a different question that feels more impressive.
Where this lesson sits in the SEC pathway
The Singapore-Cambridge Secondary Education Certificate begins in 2027. Under Full Subject-Based Banding, candidates take subjects at G1, G2 or G3 according to the level offered. The overall standard of a G3 examination corresponds to the former O-Level standard; the official 2027 G3 syllabus codes include English Language K300, Mathematics K310 and Physics K323. Scientific investigations and claims about their evidence also involve the skills described in the official G3 science syllabuses. This fictional cross-subject exercise is a practice context, not an additional SEC paper or an announced examinable case study.
SEAB lists English K300 as four papers: Writing (35%), Comprehension (35%), Listening (10%) and Oral Communication (20%). Mathematics K310 has two papers of 90 marks, each weighted 50%, and includes applying mathematics to real contexts. Students taking pure Physics K323 or a combined Science combination must use the syllabus and paper structure for the subject actually registered. Do not substitute a pure-subject syllabus code for a combined-science code.
Read the official SEC explanation, the 2027 G3 subject-code index and the relevant subject PDFs as primary references. For the immediately preceding steps, review Volume 0989 on cold-window physics, Volume 0990 on measurement and rates and Volume 0991 on reading the report.
First separate the four possible stories
Condensation occurs when water vapour changes into liquid water. Moist air meets a sufficiently cool surface and water droplets can form there. Calling droplets ‘leaking water’ is a description of their movement, not necessarily their origin. A leak through a window frame usually involves water travelling through a physical opening or joint, commonly associated with rain or another liquid-water source. These pathways can look alike once a puddle forms on the floor.
A third possibility is direct human activity, such as cleaning or an overturned container. A fourth is mixed causation: water could condense on glass while a rain-driven gap also lets water through. Good reasoning makes the alternatives explicit before deciding which tests would distinguish them. A single photograph of droplets has weak power to discriminate among these stories.
The basic physics chain for condensation is: moist air, a surface cooler than the surrounding air, and a temperature condition suitable for vapour to condense. Temperature changes alter the ability of air to hold water vapour, but students should not treat all humid air as if it must form droplets. The surface temperature, air humidity, air movement and location matter. A cold window can have condensation even when rain has not fallen.
The basic rain-entry chain is: liquid rainwater contacts a building envelope, a pathway permits entry, and water appears inside. The testable clues may include timing with rainfall, where water first appears and whether the pattern tracks wind-driven rain. None of these clues alone guarantees a leak. If a cleaner wiped the frame on a rainy morning, timing still needs careful interpretation.
Build an evidence table before calculating
Imagine a teaching case involving three observation stations: A is a cool window panel; B is an equal-sized room-temperature comparison panel; C is the bottom frame near a suspected seal. The observations are hypothetical and are provided to practise examination reasoning, not to establish anything about a real school.
Record date, start and end time, panel temperature if measured, surrounding air conditions, rain conditions, initial and final collector mass, exposed area and where the water was first observed. Give every column a unit. A picture can show the location and extent of droplets but cannot itself measure an exact mass, temperature or source. State whether the weighing includes the tray. Note who took each reading and whether the tray was moved.
Suppose panel A with a 160 cm² exposure records net collection of 6.4 g in forty minutes and panel B with the same area records 1.6 g. The average net collection rates are 0.160 g/min and 0.0400 g/min. Panel A recorded four times the net collection of B in that trial. This is a mathematical comparison, not proof that every one of A’s grams arose from condensation. A wet cloth touching A, an unnoticed spill or a leaking frame would matter.
Now add C, which shows liquid appearing at the edge during heavy rain but lacks a measured collection mass. C may warrant inspection, but it must not be assigned a made-up mass so that all three stations can be ranked. A careful report separates quantitative records from qualitative observations and identifies the next useful measurement.
Make the calculations honest
Net collected mass equals final combined mass minus initial combined mass, provided the method and container are consistent. Dividing net mass by elapsed time gives an average observed rate over the period. Dividing by exposed area yields gain per unit area; dividing again by time gives an average area-normalised rate. The denominator always determines what the result means.
For A, 6.4 g divided by 160 cm² is 0.0400 g/cm² in forty minutes. The average area-normalised rate is 6.4/(160 × 40) = 0.00100 g/(cm²·min). For B, the corresponding figures are 0.0100 g/cm² and 0.000250 g/(cm²·min). These comparisons are meaningful within the stated conditions. If the panels differ in temperature or position, those factors still limit causal interpretation.
‘Four times as much’ and ‘300% more than B’ can describe the same comparison: (6.4−1.6)/1.6 × 100% = 300%. The sentence ‘A was 300 percentage points wetter’ is not justified because the quantities are masses, not percentages on a shared scale. A badly chosen denominator can turn a correct arithmetic procedure into an inaccurate claim.
Where intervals differ, compare rates rather than totals. A 6.4 g gain after forty minutes is not directly comparable with another system’s 7.0 g after eighty minutes. Show the calculation, give a unit and return to the actual physical question. Do not infer the future rate will remain constant unless a model is explicitly assumed and defended.
Design a test that can change your mind
First observe safely without opening electrical fittings, climbing outside windows or dismantling frames. Ask a qualified person to inspect suspected building damage. In a school investigation, a student can propose recording the exact location and timing of droplets, logging indoor conditions where school equipment permits, and comparing wet-weather and dry-weather observations with supervision.
A fair test varies one chosen factor where practical, controls relevant alternatives and repeats readings. For example, keeping panel area and exposure time equal makes the mass comparison clearer. But identical areas alone do not make two positions equivalent if their air flow or surface temperature differs. A fair test is a reasoned design, not a ritual phrase.
Propose competing predictions. If condensation dominates, droplets may form on sufficiently cool glass during humid conditions even without rain. If rain entry dominates, a pattern closely linked to rain exposure and a particular frame joint may warrant further checking. A weak result may leave both possibilities open. The correct response is to improve evidence, not to push the answer toward a favoured story.
Write in three layers: observe, infer, recommend
The observation layer states what was actually seen or measured: ‘Station A had 6.4 g net collection over forty minutes in the model trial.’ The inference layer says what that might mean: ‘The colder surface is consistent with more condensation, but the source of all collected water has not been established.’ The recommendation layer says what should happen next: ‘Repeat the trial in comparable conditions and ask maintenance staff to inspect the frame if rain-linked edge wetness continues.’
A responsible school notice should never pretend that an exploratory student model certifies the safety of a classroom or building. Technical maintenance decisions belong to the relevant school staff and qualified professionals. Students can still demonstrate excellent inquiry skills by explaining limits and making sensible next-step suggestions.
Use verbs deliberately. ‘Recorded’ belongs to readings. ‘Indicates’ and ‘is consistent with’ express varying degrees of interpretation. ‘Proves’ demands strong evidence. ‘May’ is useful for genuine uncertainty, but a report full of ‘may’ without a clear conclusion is unhelpful. Tell the reader what the current evidence can support before adding the limitation.
Applied cases: test the claim, not just the arithmetic
Case 1: One photograph
A photograph shows droplets along the lower pane, and a student declares a faulty roof. The image supports the presence and location of visible moisture, not its route through the building. Ask what additional observations would distinguish cooled-glass condensation from rain entry. A picture is valuable evidence when its claim is limited to what it can show.
Case 2: A rainy morning
Rain fell before the report, so a class calls every droplet a leak. Rain is a relevant timing clue but not proof of entry. Humid air and cold surfaces can coexist with rainy weather. The investigation should record both the rainfall relationship and the location where water first appears; a single coincidence is an uncertain diagnosis.
Case 3: A dry-day repeat
Similar droplets form on cold glass on a day without rain. This strengthens the plausibility of condensation, although it does not establish the source of every previous puddle. A defensible inference is that rain is not necessary for some of the observed moisture. The older rainy-day event may still have had mixed causes.
Case 4: Temperature missing
A claim says that a window was ‘much colder’, yet no surface-temperature readings were taken. It may have felt cool to touch, but that is not a measured temperature difference. Write ‘felt cooler’ as an observation if appropriate and propose supervised measurement. Do not invent degrees Celsius merely to make the explanation look scientific.
Case 5: Humidity missing
A student reports 90% relative humidity without a hygrometer record. The number is unsupported, even if the room felt humid. A good answer can discuss humidity as an important possible factor without attaching an arbitrary figure. In an experiment, document when and where the environmental reading was made and the device used.
Case 6: The wrong baseline
A scale shows 96.2 g at the end of a trial and the report says 96.2 g of water collected. The tray might account for most of that mass. Subtract the initial tray-and-collector mass from the final total measured under the same procedure. Only the resulting net change can be compared with another collector’s net change.
Case 7: Unequal window sizes
Panel A gains 6.4 g across 160 cm²; Panel D gains 7.0 g across 280 cm² over the same period. D has a larger total gain, but A has the larger amount per area: 0.0400 versus 0.0250 g/cm². State which performance measure answers the question; ‘better’ has no secure meaning without a defined comparison.
Case 8: Unequal times
A has 4.2 g after thirty minutes; B has 4.8 g after sixty minutes. A’s observed average rate is 0.140 g/min while B’s is 0.0800 g/min. Do not rank them only by final mass. The conclusion is about the recorded average rate, not the instantaneous rate or long-term behaviour of either setup.
Case 9: Percentage denominator
A collects 6.4 g and B 1.6 g. Their absolute difference is 4.8 g; relative to B, A collected 300% more. Relative to A, B collected 75% less. Both calculations are valid with the correct wording. The error is to present a percentage without saying which measured quantity supplies the denominator.
Case 10: The zero baseline
There is no collected water at the beginning, so a pupil wants a percentage increase from zero. Dividing by zero is not defined. Report the change in grams or compare with a nonzero reference instead. A dramatic phrase such as ‘infinite improvement’ is not a meaningful interpretation of this experiment.
Case 11: The tray is moved
The initial mass was measured on one balance and the final mass on another. Different calibration offsets may affect the apparent gain. The class should record the method, check the instruments or repeat with one consistently calibrated balance. An arithmetic subtraction cannot repair an unverified measurement process.
Case 12: An unseen spill
The collector may have received water from a knocked-over cup. The reading is still the reading, but its attribution to condensation is insecure. Flag the possible contamination and repeat the controlled observation if appropriate. Deleting the result silently conceals the uncertainty and makes the subsequent explanation look stronger than it is.
Case 13: The frame joint
Water first appears at one frame joint during wind-driven rain. This points toward a possible entry pathway, but a school should ask qualified maintenance staff to check rather than relying on an unsupervised student diagnosis. A useful report describes the location precisely and notes whether dry-weather observations show the same pattern.
Case 14: Two causes together
Droplets on the middle of the glass persist on dry days, while a separate wet trail follows the frame edge during a storm. The evidence is consistent with more than one process. A learner should resist the demand to name a single exclusive cause when distinct observations support different explanations. A mixed conclusion can be the most accurate one.
Case 15: Confusing fog with liquid
A fogged pane and a puddle are treated as the same measure. Visible fogging indicates many small droplets on the surface; a puddle describes liquid collected elsewhere. They may be related, but they are not numerically interchangeable. Describe precisely what was observed and use a suitable instrument if the question requests a quantity.
Case 16: No repeat trials
Only one forty-minute run is available. It is valid to calculate that run’s gain and rate, but not to claim that its ratio will hold every morning. Repeating observations under comparable conditions can reveal variation and possible mistakes. Make the reporting period explicit and avoid universal language.
Case 17: A missing rainfall log
The report calls a morning ‘rainy’ because a pupil remembers hearing rain. Memory can be a clue, but it is weaker than a dated observation or appropriate official record. The decision might remain tentative until timing is established. Good evidence handling distinguishes witness recollection from an instrument reading without dismissing either automatically.
Case 18: Rounded readings
The scale displays mass in 0.1 g increments, while the calculation ends with nine decimal places. Extra calculator digits do not create precision in the measurement. Apply the rounding requested in the question and retain enough working figures for accurate calculation. A small observed difference should not be presented as experimentally decisive without considering measurement limits.
Case 19: Interval gradient
Net mass rises from 1.2 g at minute ten to 3.0 g at minute twenty-five. The average interval rate is (3.0−1.2)/(25−10) = 0.120 g/min. Do not divide the later total by the later time to answer a question about this particular interval. Read both coordinates before finding a gradient.
Case 20: Area-time units
Someone reports 0.00100 g/cm²/min as ‘one gram per minute’. These units mean entirely different things. For the stated 160 cm², forty-minute example, 0.00100 g/(cm²·min) is an average area-normalised rate, not a total collection rate. Multiply by area if you need the equivalent total rate under the same assumptions.
Case 21: Visual scale manipulation
A bar chart starts near the largest values and makes a small mass difference look enormous. Its bars may encode correct numbers but exaggerate their apparent contrast. Check the axis origin and labels before interpreting a graph. Where appropriate, plot net gains on a clear scale and report exact values alongside the picture.
Case 22: Correlation and cause
Colder panels tend to collect more water in three observations. This association is consistent with condensation but is not proof that temperature alone drove the differences. Location, humidity, exposure and contact with other water sources may vary. A causal investigation should account for competing variables as far as the setting permits.
Case 23: Unfairly matched positions
Panel A faces an air-conditioning outlet while panel B sits against a warm interior wall. Equal exposed area does not make them identical environments. Describe the relevant differences, then design a comparison that better controls location or measures the variables. The best conclusion may be that the existing data do not isolate one factor.
Case 24: The word ‘always’
A notice claims that droplets on indoor windows always indicate defective glazing. One counterexample—condensation on sound glass—is enough to show why this universal claim is unsafe. Replace it with a narrower statement about possible causes and useful checks. Precise qualification is a strength in scientific and English examination responses.
Case 25: Question asks ‘suggest’
A prompt says ‘suggest one reason’ and a pupil writes four unranked stories. A better answer offers one scientifically plausible explanation tied to the stated observations. When the word is ‘evaluate’, by contrast, compare alternatives and limitations. Marks are earned by responding to the command word, not by accumulating unrelated facts.
Case 26: Question asks ‘calculate’
A pupil explains humidity for a page but omits the required rate. Start with the formula, substitute observed quantities, calculate and supply units. A brief sentence can then interpret the result. The physical context tells you what the numbers represent; it does not replace numerical working when calculation is requested.
Case 27: Question asks ‘evaluate’
A student repeats that A collected more than B and calls the method reliable. Evaluation should identify a strength, a limitation and a relevant improvement. For example, equal collection time helps, but unequal panel temperatures or unrecorded spills weaken attribution. Suggest a controlled repeat that addresses the specific weakness.
Case 28: An official-sounding label
A student report is headed ‘Certified Condensation Safety Clearance’. The title implies a professional conclusion the group cannot provide. Use a description such as ‘Classroom Moisture Observation: Preliminary Student Report’. Clear naming protects readers from mistaking a learning investigation for qualified building inspection or a formal safety certification.
Case 29: The recommendation comes first
A writer begins ‘Replace all windows now’ before presenting any evidence. A more responsible structure gives the actual observation, identifies plausible interpretations and recommends targeted inspection if the concern persists. The strength of a recommendation should match the strength and urgency of evidence, while urgent safety issues go directly to responsible staff.
Case 30: An 80-word summary trap
A Paper 2-style summary task asks for particular causes and precautions, but the candidate adds an unrelated opinion on school spending. Summaries must follow the requested focus and select relevant points from the text. Keep the response near the required length, use concise paraphrase and do not manufacture facts absent from the passage.
Case 31: Over-reading the graph
The line of collected mass increases over the sample period and a student predicts it must keep increasing at the same rate forever. The graph records a finite observation window. Extrapolation requires an explicit model and may fail as conditions change. Distinguish measured trend from a conditional prediction.
Case 32: A cautious final finding
The data support higher observed moisture collection at the cooler pane, while a rain-associated edge trail remains unexplained. A strong conclusion states both findings and assigns the next action to qualified school staff. Caution does not require silence: it requires drawing the sharpest boundary the evidence permits.
Construct a complete recommendation from the evidence
A short, effective finding might begin: ‘In the illustrative forty-minute comparison, the colder panel recorded 6.4 g of net collected water, against 1.6 g at the comparison panel.’ The next sentence should say that this supports a difference in the recorded result but does not prove the origin of each gram. Add the observation about rain-associated wetness at the joint as a separate issue. Then recommend an appropriate supervised repeat and inspection of any persistent rain-linked pathway.
This sequence serves three readers at once. A science examiner can see a valid physical explanation and the limits of the evidence. A mathematics examiner can verify the quantities and units. An English examiner can follow a clear argument suited to its audience. It is stronger than three disconnected paragraphs that happen to contain subject names.
Practice set: attempt before checking
Task 1. The same collector weighs 72.3 g before and 77.1 g after a sixty-minute trial. Calculate net gain and average rate. Answer: 4.8 g and 0.0800 g/min. Explain why 77.1 g is not necessarily the collected-water mass.
Task 2. A 100 cm² panel gains 3.0 g in thirty minutes. Another 150 cm² panel gains 4.0 g in the same time. Compare their net gains per unit area. Answer: 0.0300 g/cm² and approximately 0.0267 g/cm². The first has the greater gain per area although the second has the larger total.
Task 3. Name two observations that help distinguish condensation from rain entry. Example answer: record moisture on non-rain days when the window is cool; record whether water appears first along a joint specifically during rain. Neither alone should be overclaimed.
Task 4. Repair the sentence ‘The experiment proves our classroom has no leak.’ Example answer: ‘The observations are consistent with condensation on the glass, but the investigation does not rule out water entering at the frame.’ The revised claim uses evidence and respects its limits.
Task 5. Draft a seventy-word note for a facilities team. Include the observation period, a measured result, one uncertainty and a sensible next check. Keep the language practical and avoid unsupported accusations. Compare your answer with the observe–infer–recommend structure above.
A disciplined revision routine
On the first pass, read the situation without calculating. Underline what is directly given, circle the quantities with units and mark the requested outcome. On the second pass, choose the scientific model, define the mathematical comparison and check for alternative explanations. On the third pass, write the answer in the order the question needs and remove words that claim more than the observations support.
After practice, record each error by type: misunderstood physical process, wrong baseline, wrong denominator, unlabelled unit, unfair comparison, exaggerated claim or wrong audience. Rework one example of each error the next day rather than memorising a polished answer to only one window scenario. Transfer matters: a new exam passage may concern another physical setting, but the discipline of separating observation, inference and recommendation remains useful.
Questions learners and parents ask
Will this exact classroom scenario appear in SEC?
No claim is made that it will. It is an original practice scenario designed to exercise relevant reasoning, data-handling and communication skills. The examinable content and paper requirements come from official SEAB syllabuses, not from a prediction about question topics.
Should I memorise the values in the example?
No. Practise recognising which number is the initial mass, which is the final mass, what period elapsed and which denominator the question requires. Being able to explain the measurement model is more portable than memorising 6.4 g or forty minutes.
What if I cannot determine the cause?
State the strongest observation the evidence supports, explain which competing causes remain possible and propose a specific improvement or next observation. In a question asking for a single likely cause, make a justified choice without falsely declaring certainty.
Does one weak investigation mean I cannot make any recommendation?
No. A limited investigation can justify a limited next step, such as repeat measurements or professional inspection of a persistent damp patch. It may not justify sweeping claims about all windows. Match the size of the recommendation to the evidence.
Which books or routes help with gaps?
Students moving from PSLE to secondary work can revisit Sengkang secondary transitions after PSLE, then strengthen G3 Mathematics, G3 Science and G3 English in the specific area where the reasoning first breaks.
The final check
Ask six questions before submitting: Did I say what was measured? Did I avoid treating a hypothesis as a fact? Did I use the correct baseline and units? Did I compare genuinely comparable conditions? Did I answer the command word? Is the recommendation proportionate to the evidence? A learner who can answer all six is doing more than completing one example. They are developing an examination habit that travels across English, Mathematics and Science.
Official references: SEAB SEC overview; 2027 G3 syllabus directory; K300 English Language; K310 Mathematics.
