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 1374 | EMS Greenhouse LED Independent Review — Electrical Evidence, Feasible Budgets and Public Claims

A final decision must survive independent checking

A greenhouse lighting trial is nearly complete. One student has calculated electricity consumption. Another has compared light-meter readings. A third is preparing to tell the school community that the lower-power fixtures are better. Before that announcement, a fourth person is asked to perform an independent review. The reviewer must check the physics, recompute the mathematics, ask what the experiment actually measured and edit the public claim until it matches the evidence. This volume turns earlier subject practice into a defensible decision without pretending the subjects are one examination.

The Singapore-Cambridge Secondary Education Certificate starts in 2027. There is no official combined ‘EMS examination’. G3 English Language K300, Mathematics K310 and Physics K323 are separate syllabuses; other G3 science subjects and combinations have distinct codes and assessment formats. The 2027 SEAB G3 directory is the current authority. The following independent-review exercise is wholly fictional. Its numbers do not describe a real greenhouse, installation or electrical tariff.

Review pathway: identify the decision; audit the lamp’s power and operating time; recompute an energy-only budget; separate illumination at the centre from illumination at the edge; decide whether repeats and controls are sufficient; check the safety boundary; revise a misleading written or spoken claim; and make a cautious recommendation about further supervised testing. The earlier Vol 1371 Physics guide, Vol 1372 Mathematics guide and Vol 1373 English guide build the individual skills used here.

Frame the committee’s actual question

The committee must not begin by asking which lamp looks most impressive. It needs a defined choice: under a stated monthly energy cap and a requirement to collect adequate evidence, which proposal warrants a further trial? That is different from asking which lamp should be installed permanently. Permanent adoption would require information about electrical safety, installation, durability, biological outcomes and full cost not supplied in the fictional case.

The review has three evidence layers. Physics checks power, energy and sensor measurements. Mathematics checks cost and feasible operating schedules. English checks the audience’s likely interpretation of the report. The layers do not cancel one another. A correct energy calculation cannot repair a biased light comparison; a beautifully phrased speech cannot turn a narrow result into proof of improved crop growth.

To make the exercise concrete, suppose options A and B each use four lights. A has rated power 72 W per lamp; B has 50 W per lamp. All lights would operate five hours daily for thirty days. The assumed exercise electricity rate is S$0.30 per kWh. These values are inputs to a teaching model, not commercial quotations. No equipment purchase, installation or maintenance figure is available.

Audit electrical power and energy first

Option A has combined rated power of 4 × 72 = 288 W or 0.288 kW. Its energy for five hours daily over thirty days is 0.288 × 5 × 30 = 43.2 kWh. Option B has combined rated power of 4 × 50 = 200 W or 0.200 kW. Its energy is 0.200 × 5 × 30 = 30 kWh. The monthly modelled difference is 13.2 kWh. At the assumed rate, the energy-only estimates are S$12.96 for A and S$9.00 for B, a difference of S$3.96.

This comparison is legitimate under the specified operation assumptions. It is not an estimate of the complete greenhouse bill. Pumps, fans, controllers and other expenses are absent from the data. The watts describe electrical power rather than a measured plant response. The reviewer must note whether figures are rated or measured and refrain from attaching more certainty than their sources justify.

If a trainee says B consumes 31% less electrical energy under the identical schedule, that is approximately consistent with (72 − 50)/72 × 100 = 30.56% for the lighting component. If the trainee says the greenhouse’s whole electricity bill will fall 31%, the claim changes its denominator without supporting data. Identifying that change is an evidence audit, not pedantry.

A budget changes the set of feasible choices

Suppose a fictional school limit is S$15 per thirty-day month for lighting energy only. At five hours per day, both A and B fit the cap. Option A’s modelled cost is S$12.96; B’s is S$9.00. If the schedule rises to six hours daily, A’s energy use is 51.84 kWh and its assumed charge S$15.552, which exceeds the cap. B’s six-hour energy use is 36 kWh and its assumed charge S$10.80, remaining under the cap.

A decision that includes an upper limit must respect the inequality, not simply round a calculation to a more convenient value. For A, C = 2.592h over the stated month, so C ≤ 15 gives h ≤ about 5.79 hours per day. For B, C = 1.8h, giving h ≤ 8.33 hours per day under the same rate and period. These are budget bounds, not recommendations about how long plants should be illuminated.

Mathematics has established which schedules are feasible financially within the simplified model. It has not established which hours are biologically adequate. A committee that treats an energy-cap solution as a growing plan has confused an economic constraint with a separate scientific outcome.

New evidence: the centre and the edge disagree

The earlier trial used one fixed sensor position. The independent review now proposes a fictional grid with a centre point and an edge point under controlled conditions. Suppose four repeated measurements give mean centre illuminance of 430 lux for A and 460 lux for B, but mean edge illuminance of 345 lux for A and 270 lux for B. The centre comparison favours B on that measurement; the edge comparison favours A. Neither result by itself establishes that either lamp offers more suitable light to plants.

This new evidence is deliberately designed to challenge a single-number conclusion. B’s stronger centre reading cannot be advertised as uniformly higher illumination across the greenhouse when the edge reading is lower. A’s higher edge reading cannot automatically be declared superior either. The review needs more marked positions, repeat readings and a definition of what spatial distribution is acceptable for the actual purpose.

A lux meter is weighted for human visual response. In a plant-light context, it is a limited proxy rather than a direct measure of every photosynthetically relevant quantity. The reviewer may report lux readings accurately without pretending they are direct measures of crop growth. Light distribution and biology are distinct research questions.

Evaluate controls and repeatability

A fair comparison requires the sensor at the same marked locations with stable orientation, comparable ambient light and a consistent reading schedule. The trial should document instrument settings and any abnormal observations. Four repeats at a centre and edge are a starting point for analysis, not proof that the behaviour is universal. A change in sunlight or distance could easily distort a local comparison.

The reviewer should look for both random variation and systematic bias. Scatter between repeated readings suggests a need to describe variability. A sensor consistently shifted closer to one lamp introduces a persistent geometry problem that averaging cannot repair. An anomalous reading should be recorded and investigated rather than silently removed. Every suggested improvement should name the flaw it addresses.

The team must also say what has not been tested: long-term operation, actual electrical installation, device durability and plant-growth response. Making that list protects the public conclusion against unsupported expansion. It is not a demand to measure everything in one class; it helps the school choose the next test that would genuinely reduce uncertainty.

Check the safety boundary before any recommendation

Electricity near water introduces hazards beyond arithmetic. Damaged insulation, damp equipment, overheating and unsuitable electrical connections require appropriate precautions. The Physics K323 syllabus covers electricity safety, but syllabus knowledge is not permission for unsupervised installation. Classroom investigations should use teacher-approved safe apparatus and adult supervision. Students should never handle exposed mains wiring or bypass protective devices.

An independent reviewer should separate ‘the energy model looks feasible’ from ‘the equipment is safe to install’. The latter requires qualified assessment and appropriate installation standards not contained in the fictional calculations. If a plug or cable appears damaged, stop the demonstration and notify the responsible adult. The review cannot approve a real electrical installation based on a school model.

A safe recommendation may be to gather further controlled measurements in a supervised educational setting, not to modify a greenhouse immediately. Safety is a constraint that cannot be traded away because another metric looks favourable.

Rewrite the committee’s public claim

The proposed announcement reads: ‘The new LED lights are proven to make greenhouse plants grow faster while cutting all power costs by 31%.’ It combines two unsupported claims. The study has no plant-growth measurements, and a percentage based on lamp wattage cannot be applied to the entire electricity bill. The centre and edge readings also differ in opposite directions, so a single ‘brighter everywhere’ assertion would be unjustified.

A corrected public summary might say: ‘Under the assumed operating schedule and energy price, option B has a lower modelled lighting electricity cost. Our small classroom readings differed by measurement location: B was higher at the centre, while A was higher at the edge. We have not tested plant growth or verified installation suitability. Further supervised comparisons are needed before making a real-world equipment decision.’

Good English preserves the evidence’s limits without losing a clear message. The audience should know what was calculated, what was observed and what remains undecided. A recommendation becomes more credible when it can be changed if new evidence appears.

A cross-subject answer is not an official combined paper

English, Mathematics and Physics have different examination components. The SEC does not have a special combined EMS paper. The integrated case helps a student practise a transferable chain of reasoning, but actual candidates must follow their enrolled subject’s current syllabus. The SEAB directory distinguishes G3 Pure Science and the relevant combined Science syllabuses; do not assume K323 applies unchanged to all Science students.

In Mathematics, show equations, conversions, inequalities and contextual interpretations. In Physics, describe quantities, measurement conditions, experimental controls and safety. In English, answer the speaking or writing task for its audience and maintain accurate qualifications. These are complementary skills, not interchangeable mark schemes.

An integrated practice task can expose a hidden weakness. A student may calculate accurately but write an exaggerated conclusion; another may speak fluently but misread a percentage; a third may know scientific terminology without designing a fair test. The next revision step should repair whichever link failed rather than assign more of the subject the learner already handles well.

How families and learners can use this audit

Ask the student to explain the numbers without looking at notes: why do the two lamp sets use different energy; why does one exceed the six-hour budget; why do centre and edge readings lead to different observations? Then ask for a one-paragraph public summary that does not promise plant growth. If the pupil can answer those questions clearly, the learning has moved beyond repeating a memorised worked example.

For supporting routes, the PSLE learning guide helps repair foundations in measurement and number sense. The Mathematics owner, Science laboratory reporting owner and English practice owner offer subject-specific preparation.

The best final decision in this fictional case is not a winner announced by an attractive statistic. It is a conditional recommendation to review more evidence safely and transparently. That conclusion differs from the earlier root-tissue audit because the current problem concerns electrical energy, spatial illumination and operating constraints. The habit that persists is independent checking before a public claim.

Thirty independent-review exercises with worked answers

These original exercises are not SEAB examination questions. They are a way to practise the same reasoning on different parts of one fictional decision. Try each task first, then compare your reasoning with the explanation. The most useful response states the quantity or claim, shows the evidence and identifies the boundary of the conclusion.

Case 1: Identify the actual decision

Task. The school committee wants to choose a lamp today. What question should its reviewer ask first?

Reasoned answer. The reviewer should ask whether the decision is about electricity cost, measured illumination, plant response, safe installation or a combined requirement. The classroom information addresses some of these questions but not all. Setting the decision criterion before calculating prevents a neat answer to the wrong problem. A pilot comparison may be justified even when the evidence is insufficient to approve permanent installation.

Case 2: Recompute total power

Task. Four option A lamps each have rated power 72 W. Four B lamps each have 50 W. Calculate combined power.

Reasoned answer. For A, four times 72 W equals 288 W, or 0.288 kW. For B, four times 50 W equals 200 W, or 0.200 kW. The labels describe nominal power under the stated model, not total monthly energy. If the lamps run at different duty cycles, the subsequent energy calculation must reflect that rather than use a single shared time assumption.

Case 3: The five-hour energy comparison

Task. Find monthly electrical energy when both plans run five hours daily for thirty days.

Reasoned answer. A uses 0.288 × 5 × 30 = 43.2 kWh. B uses 0.200 × 5 × 30 = 30.0 kWh. The modelled difference is 13.2 kWh. These figures use consistent running time and rated power. They do not include pumps, cooling, controllers or any other electrical loads. The reviewer should state that boundary before translating an energy figure into a wider operating claim.

Case 4: Assumed monthly costs

Task. At a fictional price of S$0.30 per kWh, compare the electricity-only cost of A and B.

Reasoned answer. For A, 43.2 × 0.30 = S$12.96. For B, 30 × 0.30 = S$9.00. The difference is S$3.96 per month under the particular schedule and price. No current Singapore electricity tariff is asserted. A report that omits the assumed rate may encourage readers to mistake an exercise estimate for a live quotation or a guaranteed saving.

Case 5: The six-hour budget

Task. At six hours daily, does A satisfy a monthly lighting-energy cap of S$15?

Reasoned answer. No. Its energy is 0.288 × 6 × 30 = 51.84 kWh. At the assumed rate, the cost is S$15.552. That exceeds the S$15 cap by S$0.552. The correct conclusion is that the proposed six-hour A schedule is financially infeasible in the simplified model. Rounding a cost down to S$15 would obscure the violation rather than solve the inequality.

Case 6: An alternative under the cap

Task. Does B at six hours daily satisfy the same S$15 energy-only cap?

Reasoned answer. Yes, under the given assumptions. The monthly energy is 0.200 × 6 × 30 = 36 kWh, and its estimated energy-only cost is S$10.80. Feasibility under the cost cap does not establish that the six-hour lighting schedule supports any intended plant outcome. Report ‘within the energy budget’ instead of ‘approved for growing’.

Case 7: A’s maximum hours

Task. Find the maximum average daily operating hours for A at the assumed tariff and S$15 cap.

Reasoned answer. The monthly cost is 0.288 × h × 30 × 0.30 = 2.592h. Solve 2.592h ≤ 15, giving h ≤ about 5.787 hours. If the plan requires an integer number of daily hours, five is the maximum permitted. The answer is an upper bound from cost, not a target duration recommended by any plant-science source.

Case 8: B’s maximum hours

Task. Find the corresponding modelled upper bound for B under the same cap.

Reasoned answer. B’s monthly energy-only cost is 0.200 × h × 30 × 0.30 = 1.8h. Thus 1.8h ≤ 15 gives h ≤ 8.333 hours per day. This larger budget ceiling results from lower rated power, but operating more hours may change outcomes not examined by the cost model. The reviewer should keep a financial ceiling separate from a safe or useful biological schedule.

Case 9: The percentage trap

Task. A headline says ‘New lamps reduce the whole greenhouse bill by 31%’. Is that supported?

Reasoned answer. No. The 22 W difference per lamp is about 30.56% of the original 72 W rating. Under identical count and operating schedule, that is approximately the relative reduction for the lighting component’s modelled electrical energy. The entire greenhouse bill may include other loads and charges. The headline silently changes the comparison base and should be corrected.

Case 10: The missing equipment price

Task. A purchaser claims that B will repay its extra purchase price in six months. Can the reviewer verify that?

Reasoned answer. No purchase-price difference has been supplied, nor information about maintenance, installation or lifetime. The reviewer can compute the energy-only modelled saving but cannot compute payback from that alone. A correct response identifies the missing information rather than inventing an equipment price. A financially plausible claim is not necessarily a supported one.

Case 11: One sensor is not a greenhouse

Task. B has a higher mean reading at the central sensor. Can the team call it brighter everywhere?

Reasoned answer. No. A central measurement reports conditions at one point. The hypothetical second-stage measurements show A has a higher reading at the edge. A claim of uniform superiority would contradict those observations. Map several fixed positions with repeated measurements before describing distribution. Even a reliable map would not automatically establish the crop’s biological response.

Case 12: Compare centre readings

Task. The mean centre readings are A 430 lux and B 460 lux. Calculate and interpret the difference.

Reasoned answer. At the centre, B exceeds A by 30 lux in the fictional comparison. A valid statement is that B produced the higher mean centre sensor reading under the tested conditions. This is not evidence that B uses thirty more watts, nor that plant growth rises by thirty units. The measured variable, location and unit must remain attached to the conclusion.

Case 13: Compare edge readings

Task. At the edge, A averages 345 lux and B averages 270 lux. What differs?

Reasoned answer. A exceeds B by 75 lux at the edge in the fictional comparison. This reverses the rank seen at the centre. The review should not select a preferred source from either number without deciding what distribution is required and how adequate coverage will be measured. Conflicting local observations can be informative because they reveal that a single measurement position was insufficient.

Case 14: An unjustified overall average

Task. A student averages centre and edge values and calls the result the average light received by all plants. Identify the problem.

Reasoned answer. Two selected measurement positions are not necessarily a representative sample of the growing area. The calculation could be an arithmetic mean of two numbers, but labelling it an area-wide average assumes equal representation without evidence. A systematic grid with defined positions and suitable treatment of coverage is needed before making a spatial claim. Arithmetic correctness does not repair a sampling flaw.

Case 15: A sensor’s scope

Task. Does a lux reading directly measure photosynthetically useful light or future crop mass?

Reasoned answer. No. Lux is weighted for human vision. It can be used to describe illuminance at a sensor position, but is not itself a complete plant-light measure or a biological growth result. A separate instrument and biological investigation may be required for those questions. The reviewer should not promote an accessible proxy into a universal response variable.

Case 16: A daylight confounder

Task. The B edge readings were taken after afternoon clouds arrived. How should the team proceed?

Reasoned answer. Changing ambient daylight may have contributed to the lower B readings. The study should compare like conditions, record or control background illumination and repeat the relevant measurements. The reviewer should mark the existing comparison as provisional rather than erase the readings. A specific cause-and-repair explanation is stronger than merely saying the experiment was ‘not fair’.

Case 17: A systematic placement error

Task. What happens if B was always measured farther from the sensor at the edge?

Reasoned answer. The geometry could lower B readings independently of lamp characteristics. Averaging more readings taken under that same biased setup would not isolate the lamp effect. Use a fixed stand and measured positions for both sources, then repeat. The review should distinguish the quality of the measurement procedure from the number of observations collected.

Case 18: Repeatability versus representativeness

Task. Four readings are stable at one location. Does that demonstrate adequate sampling across a greenhouse?

Reasoned answer. No. Stability indicates limited within-location repeatability, not representation across all positions or times. A larger spatial grid and observations under varying relevant conditions may be needed. Repetition answers how consistent a measurement is under a stated setup; sampling design answers which wider conditions the result represents. These are separate dimensions of evidence quality.

Case 19: An outlying result

Task. A single edge reading is far higher than the others. What is an honest reporting process?

Reasoned answer. Preserve the original value, check its recording and the measurement setup, document any reflection or sensor movement, and repeat under the specified protocol. Do not drop an observation because it makes the preferred lamp look less convincing. Report the investigated issue and any replacement measurement clearly. A committee should be able to reconstruct the evidence trail.

Case 20: The safety veto

Task. An equipment cable is visibly damaged near water. Can low calculated energy cost justify proceeding?

Reasoned answer. No. A damaged electrical connection in a damp environment is a hazard, not a financial trade-off. Stop the demonstration and inform the responsible adult. Students must not inspect exposed mains conductors or improvise repairs. The independent review should record that the exercise’s numerical feasibility does not confer permission to use unsafe apparatus or make a real electrical installation.

Case 21: A deceptive headline

Task. Rewrite: ‘Proven brighter across the greenhouse and guaranteed to grow more vegetables.’

Reasoned answer. A supported statement is: ‘The classroom sensor readings varied by position, and we have not measured plant growth.’ If the audience needs a recommendation, add that more controlled mapping and biological testing would be needed before a wider decision. The rewrite removes unsupported universal claims while preserving useful information. Accuracy, not dramatic wording, is the goal.

Case 22: A concise public cost sentence

Task. Write one sentence about the B schedule without implying an actual utility tariff.

Reasoned answer. ‘With four 50 W lamps running five hours each day for thirty days, our classroom model estimates 30 kWh of lighting energy and S$9.00 at an assumed S$0.30 per kWh.’ That sentence supplies the count, rating, time, energy and pricing assumption. A reader can reconstruct the number and see that it is not a verified real bill.

Case 23: A visitor’s sceptical question

Task. A visitor asks why the team is delaying its decision if B costs less to run. How might a speaker reply?

Reasoned answer. ‘The energy calculation favours B under our assumed schedule, but our light measurements differ by position. We still need to check coverage, safety and the plants’ actual needs before adopting equipment permanently.’ This directly answers the visitor rather than dismissing the cost evidence. The speaker acknowledges what is known and names the remaining decisions.

Case 24: Separate three columns

Task. Design a compact review table to avoid mixing facts with assumptions and claims.

Reasoned answer. Use columns such as item, evidence source, calculation or observation, assumption or limitation, and action. The energy row contains rated power, duration and modelled cost. The illumination row contains sensor location, repeated readings and geometry controls. The public claim row contains exact wording and whether it follows from the first two. This structure shows where a conclusion needs additional information.

Case 25: Conclusions and caveats

Task. Why is ‘B may be worth a further controlled trial’ different from ‘B is the winning option’?

Reasoned answer. The first statement proposes evidence gathering consistent with a lower modelled energy charge and incomplete spatial readings. The second implies a complete comparative judgement that the current measurements cannot justify. A conditional recommendation is not vague when the condition is specific. It tells decision-makers what further facts would make a stronger conclusion possible.

Case 26: No official EMS paper

Task. What should a G3 student understand about this integrated exercise?

Reasoned answer. It is a study method using English, Mathematics and Physics together, not a standalone Singapore-Cambridge SEC paper. The student must check the exact subjects enrolled and their individual 2027 syllabuses. For example, G3 Combined Science has its own subject codes and formats; the separate Physics K323 practical pattern does not automatically apply to those combinations. Cross-disciplinary learning supports transfer without replacing official preparation.

Case 27: A follow-up about priorities

Task. The committee cannot afford every possible investigation. Which test comes next?

Reasoned answer. Choose a test that most directly reduces uncertainty in the decision. Since the centre and edge results differ, controlled mapping of multiple positions would help evaluate coverage. If the intended final outcome is plant growth, a separate appropriately designed biological trial becomes necessary. Safety checks remain essential rather than optional. The exact order depends on the committee’s stated decision criterion.

Case 28: A final five-sentence review

Task. How could a student assemble an entire decision report in five sentences?

Reasoned answer. First state the decision and its assumed boundary. Second give the verified electricity comparison. Third describe the centre-edge conflict in the sensor evidence. Fourth identify what is not yet tested, including biological performance and installation suitability. Fifth recommend a supervised, controlled follow-up rather than unconditional adoption. The format forces the student to distinguish calculation, observation, uncertainty and action.

Case 29: Error diagnosis after marking

Task. A student calculates costs correctly but says B must be best. What is the learning gap?

Reasoned answer. The principal weakness is interpretation, not arithmetic. Ask the learner to list which outcomes the energy model includes and which it omits. Then present the conflicting centre-edge measurements and request a revised conclusion. Repeating more multiplication problems will not address a habit of turning narrow evidence into a universal recommendation.

Case 30: Transfer away from lamps

Task. Could the same audit structure help evaluate a school water-saving device?

Reasoned answer. Yes, with the relevant quantities changed. Instead of watts and lux, a study might compare water volumes, rates, cost and service performance. The reviewer would still define the decision, check units and calculations, control conditions, preserve anomalies and edit claims to match the measured outcomes. Transfer is demonstrated when the student adapts the evidence method instead of copying the greenhouse details.

Seven further habits for independent performance

An evidence chain has distinct links

An independent reviewer can mark each link as checked, provisional or unsupported. The power ratings are stipulated inputs, and the energy arithmetic can be independently recalculated. The centre and edge readings are fictional observations under described conditions; their validity depends on controls and repeats. Plant-growth and real installation claims remain unsupported. This approach shows why one accurate result cannot automatically upgrade every connected statement to proven truth.

Repairing an early mistake before it spreads

If a pupil accidentally treats 72 W as 72 kW, every downstream energy and cost figure becomes vastly inflated. The first repair is the unit conversion, not the conclusion. If the units are correct but the wrong time period is used, fix the model’s duration. If the arithmetic is correct but the public statement says ‘all energy costs’, fix its scope. Locating the earliest weak link reduces unnecessary rewriting and improves the student’s reasoning efficiency.

Use a decision table as a reading comprehension exercise

Ask learners to explain each row in ordinary language before drawing a conclusion. They should say what an input means, what the calculation assumes and whether the value is measured or hypothetical. Such reading discipline helps with English comprehension as much as mathematics. A persuasive heading or graph should not distract from an omitted denominator, a missing unit or a comparison made under different test conditions.

The danger of selecting only convenient results

A committee hoping to choose B might quote the lower energy charge and stronger centre reading while ignoring its weaker edge reading. A committee favouring A could do the opposite. Scientific reporting requires representing relevant evidence even when it complicates the preferred story. This does not mean every piece of information has equal quality. Explain which results are well controlled, which are provisional and why the next test is needed.

Improving the next experiment without excess complexity

For spatial light comparison, mark a small set of positions, use the same sensor and orientation, control or record ambient conditions and collect repeated readings at each point. Keep the table transparent and document abnormal results. This is a focused improvement to a specific uncertainty. Adding unrelated measurements would not necessarily strengthen the comparison. Good investigations are built around clear questions, not the largest collection of data.

A public audience deserves the calculation’s boundary

Readers may assume a number in an article describes real-world conditions unless the author says otherwise. This guide uses an invented S$0.30 rate, fixed operating hours and rated lamp powers. That boundary should accompany a quoted cost, especially if it is detached from the rest of the explanation. A technically correct figure becomes misleading when the condition that gives it meaning is quietly removed.

From separate skills to independent judgement

The learner has now progressed through an electrical investigation, an algebraic budget and an oral communication task. The integrated review asks whether those skills still work when the evidence is mixed. The strongest answer may be ‘the calculation is correct, the coverage is unresolved and the public claim must be narrowed’. That is an intellectually complete answer because it gives a clear account of what is known and what must happen next.

Conclude the lighting sequence and prepare for unfamiliar questions

This independent review completes the four-volume LED-lighting progression: Physics and measurement, Mathematics and constraints, English spoken explanation, and this integrated evidence review. Before leaving the unit, try an unfamiliar situation such as a school ventilation decision: define the measured variable, verify an energy model, inspect the sample, correct the claim and explain what further information is needed. A student who can transfer that sequence has learned more than the details of one greenhouse.