The final question is not whether the experiment looks successful
A school science club has crystals in a dish, a mathematical prediction and a report waiting to be shown at an exhibition. Three students describe the same outcome differently. One says the experiment worked because crystals appeared. Another says it failed because less mass was recovered than expected. A third calls the sample pure because it looks clean. Each speaker may be using a different definition of success. Before anyone prints the poster, the group needs a disciplined decision.
This guide is about making that decision. It connects Volume 1029 on Chemistry crystallisation, Volume 1030 on mathematical recovery and Volume 1031 on reporting. Rather than retelling those lessons, we will ask a harder question: which new observation would actually settle a dispute?
Where this approach belongs in the 2027 SEC
The first SEC examination is in 2027. Its G3 English Language subject is K300, Mathematics is K310 and Chemistry is K324. The combined science options are K326 Science (Physics, Chemistry), K327 Science (Physics, Biology) and K328 Science (Chemistry, Biology). A school candidate’s actual component requirements depend on which subjects the student enters for. These codes are from SEAB’s 2027 school-candidate list, not a continuation of the older 1184, 4052 or 6092 labels.
The practical and communication work in this article is an integrated study exercise, not a new stand-alone EMS examination paper. English Paper 1 includes 250–350-word situational writing; G3 Mathematics Paper 2 includes a real-world application question; G3 Chemistry assesses scientific investigation and explanation in written and practical papers. These assessments are distinct. Use the integrated exercise to build transferable reasoning, then practise the official demands in their proper formats. See SEAB’s 2027 G3 subject list.
Three meanings of success
Success of the process means that crystallisation occurred: solid crystals formed from a solution under the observed conditions. Recovery concerns how much solid was collected relative to a meaningful comparison. Purity concerns what substances the collected sample contains. All three can be true, partly supported or unknown at once. A large recovered mass is not automatically evidence of high purity; it could include retained water or other material.
A fourth concern is repeatability: will comparable trials produce broadly consistent observations? Without repeats or a measurement standard, a single result may be a useful beginning but not a dependable public claim. In science, failure to answer one question does not cancel the value of a properly answered different question.
Start with a claim, then ask what would make it true
Imagine the claim ‘our recovery was 82%’. What is the numerator? The observed 16.4 g. What is the denominator? A theoretical 20.0 g based on the provided solubility model. Is the percentage calculated correctly? Yes: 16.4/20.0 × 100 = 82. But the calculation’s accuracy does not prove its inputs are valid. Was the collection mass recorded after drying? Was the model’s water mass comparable to the trial? Was the starting solution saturated?
Now imagine the claim ‘our crystals are pure’. The evidence needed is different. You cannot obtain it by rearranging a mass ratio. You need an appropriate independent method to assess composition or purity, selected with a teacher’s guidance and suitable for the substance. The answer to one question should never be smuggled into another.
Read the evidence board before calculating
Here is our entirely fictional board. A solubility reference predicts 20.0 g of solid could separate under specified ideal conditions. Trial A records 16.4 g collected, with the note ‘sample appeared damp’. Trial B records 17.8 g, with the note ‘some solid remained in the vessel’. Trial C records 15.9 g but the cooling temperature is missing. A student poster describes B as ‘the purest’. These records contain useful observations and serious gaps.
The nominal mass recoveries are A: 82%, B: 89% and C: 79.5%. Those values compare the recorded masses to the same assumed ideal amount. They do not rank purity. They are also not a controlled comparison of practical efficiency if the mass condition and cooling procedure varied. The correct next move is to identify what the club needs to know rather than rush towards the largest number.
Four distinct layers in an evidence chain
The first layer is a direct observation: a mass reading, a temperature, a visible change or a recorded time. The second is a calculated quantity such as a percentage. The third is a model-based interpretation: what the quantity would mean if the assumed solubilities and starting conditions hold. The fourth is a proposed public statement or decision. Each layer needs a short explanation of how it depends on the one before.
Most bad reports jump from the first or second layer directly to the fourth. A measured mass becomes a purity guarantee. A calculated difference becomes proof that an operator made an error. An attractive photograph becomes evidence of a chemical identity. Strong students learn to pause at the missing layer and say precisely what the record does and does not establish.
Create a claim-evidence-warrant test
Write the proposed claim in one sentence. Beneath it, name the evidence actually available. Then state the warrant: the reasoning that connects that evidence to the claim. For ‘crystals formed’, an observation of crystal formation supports the claim under the recorded conditions. For ‘the collection was 82% of the model maximum’, the warrant is the correct percentage calculation using a valid comparison mass. For ‘the crystals were 82% pure’, no relevant warrant has been supplied.
This three-part structure helps both written Science answers and English evaluations. It does not mean every official answer must have labelled boxes. The labels are a training technique that makes invisible reasoning visible. Eventually the student can express the same logic as a compact sentence.
What an honest calculation can and cannot do
A calculation can reveal an inconsistency even when it cannot diagnose the cause. Suppose a group claims 21.0 g of dry recovered solid while the scenario’s theoretical maximum is 20.0 g. The nominal percentage is 105%. That is a warning to inspect the model, units, initial conditions, sample drying, transfer of unrelated material and measurement records. It is not automatically proof of dishonesty or a demonstration that conservation of mass has failed.
If a result appears physically incompatible with the model, the first question is which assumptions are wrong or missing. The model may describe only one substance in one solvent mass at two temperatures. An observed sample may include something the model excludes. Scientific caution begins with a list of testable possibilities.
Design the next observation so it can change the decision
A retest should address an identified uncertainty. If dampness is the concern, the next action is a consistent, supervised procedure for obtaining a sample mass suitable for comparison, followed by a recorded reweighing. If transfer loss is the concern, examine the collection process and record residues. If purity is the concern, choose an appropriate analysis rather than simply repeat the same mass measurement. Different questions need different tests.
The critical habit is to state what outcome would change your mind. For example: ‘If two comparable dry-mass measurements agree within the method’s precision, confidence in the recorded mass increases.’ Do not write ‘the test will prove it is pure’ unless the test can actually assess the claimed property.
The English contribution: make the decision readable
An evidence board is not an excuse to bury the audience in raw numbers. A concise report might say: ‘Trial B had the highest recorded mass, but the groups used insufficiently documented drying and transfer procedures. The data do not establish which sample is purest. We recommend repeating the trial under comparable conditions before displaying a numerical claim.’ That statement moves from observation to qualification to action.
Notice the distinction between clear and timid language. ‘We do not know anything’ would be too weak: the group knows that crystals formed and knows the recorded masses. ‘The results prove B is best’ would be too strong. The right sentence makes the supported part definite and the unsupported part explicit. Precision is not reluctance; it is competence.
The Mathematics contribution: watch the denominator
If a student uses recovered mass divided by starting mass of solute, that percentage answers ‘what portion of the initial solid was collected?’ It does not answer ‘what proportion of the theoretically crystallisable mass was collected?’ The values may be very different. In our fictional case, if 32.0 g was initially dissolved, 16.4/32.0 × 100 = 51.25%; 16.4/20.0 × 100 = 82%. Both arithmetic expressions may be correct, but only one fits the stated recovery definition.
A mathematical answer should therefore name the denominator in words before doing the calculation. The G3 Mathematics syllabus emphasises mathematical reasoning, interpretation and communication as well as techniques. Showing the calculation without identifying the quantity can leave a logically incomplete answer.
The Chemistry contribution: separate processes from outcomes
Crystallisation can separate a dissolved solid by changing the conditions that govern how much remains dissolved. Filtration can then collect the solid from the liquid. This says what the procedure is designed to do. It does not guarantee that every impurity has been removed or that the recovered solid is dry. In different circumstances, impurities may remain in the liquid or contaminate the solid; that must be evaluated rather than presumed.
The K324 syllabus includes methods of purification and analysis, as well as planning, observation and evaluation in practical work. The skill is choosing a suitable method for a particular mixture and explaining why it is suitable. Memorising the word ‘crystallisation’ without understanding its limits prepares a student poorly for unfamiliar questions.
A one-page audit template
Begin with a question and proposed claim. Then state the source and date of each observation if supplied. Add the measurement conditions, model assumptions, calculation, possible alternative explanations, decision and targeted follow-up. The format need not be a literal table. In an oral discussion, you can follow the same order in six short sentences. In a practical exercise, it can be a planning checklist before touching apparatus.
This template is deliberately narrower than a general ‘be critical’ instruction. The student must name an exact quantity, an exact missing condition and an exact next observation. When the audit becomes habitual, a complex case can be reduced to manageable decisions rather than a pile of memorised slogans.
Do not manufacture certainty with repeated trials
Three trials can be more informative than one, but repeats do not automatically remove a shared mistake. If every group weighs wet crystals using the same unsuitable method, the three measurements may agree closely yet all be biased. Precision describes repeat agreement; accuracy concerns closeness to a suitable reference. The two are connected but not interchangeable.
Likewise, an average should not conceal the range. For illustrative masses 16.4, 17.8 and 15.9 g, the mean is 16.7 g; the range is 1.9 g. Those figures are mathematically correct, but without controlled procedures they do not isolate causes or justify a purity claim. State what averaging reveals and what it cannot solve.
What to do when the science and prose disagree
Suppose a draft report says the club lost 3.6 g, and the calculation shows 20.0 – 16.4 = 3.6 g. The subtraction is sound, but ‘lost’ may be misleading. The 20.0 g is a model-based amount that could crystallise, not necessarily a measured amount that disappeared. A better phrase is ‘3.6 g below the model maximum’. The wording respects the status of the reference.
This is the special value of studying English, Mathematics and Science together. Science checks mechanism and evidence. Mathematics checks relationships and magnitudes. English checks what a public reader will understand. Each subject prevents a different kind of error.
A science-fair decision, not a sales pitch
If a display board is going to say ‘Our method gives 90% pure crystals’, the team needs to ask who might rely on that statement and what test supports it. A school poster does not become exempt from evidence because its purpose is educational. It may be perfectly appropriate to describe how crystallisation works and show the actual trial readings while saying that purity was not independently established.
The aim is not to make the exhibition less interesting. A board that explains why a large recovered mass does not guarantee pure crystals can teach visitors something deeper than a bright photograph. Intellectual curiosity grows when students learn which questions an experiment has not yet answered.
Plan a fair repeat without improvising hazards
Before repeating a supervised school practical, agree on the intended variable and the outcome being compared. Keep the starting material, solvent amount, temperature plan, collection method and mass conditions as consistent as the investigation requires. Record anomalies immediately rather than rewriting the results after a disappointing outcome. Use authorised school laboratory procedures and the teacher’s safety instructions.
This is a reasoning guide, not a step-by-step invitation to heat unknown substances at home. In a practical examination, candidates follow provided instructions and laboratory safety standards. A good planning answer identifies risks and suitable precautions in relation to the given materials without pretending that a generic recipe is safe for every chemical.
Practical and examination relevance without invented components
K324 Chemistry includes a 40-mark multiple-choice Paper 1 (30%), an 80-mark structured/free-response Paper 2 (50%) and a 40-mark practical Paper 3 (20%). The practical includes planning, manipulation, data presentation and evaluation. K310 Mathematics has two equally weighted 90-mark papers of 2 hours 15 minutes each. K300 English has its own four papers, including report-suitable situational writing under Paper 1.
These facts help students place the integrated exercise in their actual revision plans. The audit is not an additional examination format, nor does it replace practising chemistry apparatus or English oral skills. Different G3 Science combinations have different papers and specifications. Students must use their own registered syllabus, not an assumed one.
A deliberate four-session revision sequence
In session one, practise classifying statements as observations, calculations, assumptions, interpretations or recommendations. In session two, perform the same classification on a deliberately untidy data board containing a missing temperature and a contradictory note. In session three, propose a targeted retest for each missing piece of evidence. In session four, write a 250–350-word situational report for a named audience using the revised claims.
After the four sessions, ask a learner to explain one claim without notes. If the answer uses the correct denominator, distinguishes recovery from purity and names the missing test, the reasoning has become transferable. If it collapses into a memorised phrase, return to the observation and rebuild the chain. This is progression from basic classification to decisions under uncertainty.
Twenty-four decision drills with worked reasoning
Each independent drill presents a claim from a fictional school science club. Decide whether it is supported, conditionally supported or unsupported, then compare the response. These are original training situations, not SEC past-paper items.
Decision 1: Crystals appear
Proposed statement: We see crystals after cooling, so crystallisation occurred. Reasoned response: Supported in the limited sense that crystals formed under the observed conditions. It does not establish the chemical identity, purity or recovered dry mass. Record the conditions before using the observation for a wider claim. Transfer question: What single new observation, controlled comparison or source check would most improve confidence here? Record an answer in plain words before turning to the next case.
Decision 2: Mass is 16.4 g
Proposed statement: A balance records 16.4 g after collection. Reasoned response: Conditionally supported as a recorded mass; confidence depends on the weighing process, calibration, and what was on the pan. Do not silently rename it ‘dry solid mass’ if the sample was damp. Transfer question: What single new observation, controlled comparison or source check would most improve confidence here? Record an answer in plain words before turning to the next case.
Decision 3: Model predicts 20.0 g
Proposed statement: The table says 20.0 g could crystallise. Reasoned response: The prediction depends on starting saturation, water mass and the stated cooling conditions. It is an ideal comparison under those assumptions, not a guarantee that the real group will collect exactly that amount. Transfer question: What single new observation, controlled comparison or source check would most improve confidence here? Record an answer in plain words before turning to the next case.
Decision 4: A nominal percentage
Proposed statement: 16.4 g is 82% of 20.0 g. Reasoned response: Arithmetically supported: 16.4/20.0 × 100 = 82. Call it a nominal model-based recovery if the underlying masses differ in reliability. Explain the denominator rather than quoting 82% alone. Transfer question: What single new observation, controlled comparison or source check would most improve confidence here? Record an answer in plain words before turning to the next case.
Decision 5: An apparent improvement
Proposed statement: Trial B has 17.8 g while A has 16.4 g. Reasoned response: B has the higher recorded mass by 1.4 g. It does not follow that B had better chemical purity or a superior procedure; verify that both measurements represent comparable samples and collection conditions. Transfer question: What single new observation, controlled comparison or source check would most improve confidence here? Record an answer in plain words before turning to the next case.
Decision 6: A visually attractive result
Proposed statement: The crystals are white and well shaped. Reasoned response: This observation can be reported as appearance. It cannot identify a chemical compound or support numerical purity without appropriate testing. Form a separate question about identity or composition. Transfer question: What single new observation, controlled comparison or source check would most improve confidence here? Record an answer in plain words before turning to the next case.
Decision 7: A missing temperature
Proposed statement: Trial C has no cooling endpoint recorded. Reasoned response: The nominal recovery can be calculated, but comparison to an ideal model that depends on final temperature is uncertain. Ask for an accurate temperature record in the next trial rather than guessing its value. Transfer question: What single new observation, controlled comparison or source check would most improve confidence here? Record an answer in plain words before turning to the next case.
Decision 8: A mass above model maximum
Proposed statement: The collected mass is 21.0 g against a 20.0 g model maximum. Reasoned response: The nominal ratio is 105%, which prompts investigation of wet crystals, impurities, incorrect assumptions or records. Do not call it a 105% purity; the categories have changed. Transfer question: What single new observation, controlled comparison or source check would most improve confidence here? Record an answer in plain words before turning to the next case.
Decision 9: A repeated reading
Proposed statement: Two balance readings agree to one decimal place. Reasoned response: Agreement suggests repeatability at the shown resolution but not necessarily accurate dry mass. If both include moisture, repetition alone cannot settle that bias. Transfer question: What single new observation, controlled comparison or source check would most improve confidence here? Record an answer in plain words before turning to the next case.
Decision 10: A claimed loss
Proposed statement: The report says 3.6 g was lost. Reasoned response: 3.6 g is the difference between model maximum and recorded mass. That is not necessarily a physically measured loss. Use ‘below the predicted maximum’ unless the investigation identifies actual losses. Transfer question: What single new observation, controlled comparison or source check would most improve confidence here? Record an answer in plain words before turning to the next case.
Decision 11: The wrong denominator
Proposed statement: A student divides 16.4 by the initial 32.0 g. Reasoned response: The answer 51.25% describes the collected share of initial solute. It is not the 82% share of the theoretically crystallisable amount. Name the intended comparison before selecting the formula. Transfer question: What single new observation, controlled comparison or source check would most improve confidence here? Record an answer in plain words before turning to the next case.
Decision 12: One successful trial
Proposed statement: The poster says the method always works. Reasoned response: A single observed trial does not justify always. A defensible statement describes the trial and proposes further tests for repeatability. Transfer question: What single new observation, controlled comparison or source check would most improve confidence here? Record an answer in plain words before turning to the next case.
Decision 13: An unverified purity tag
Proposed statement: The sample is labelled 95% pure because the mass seems high. Reasoned response: Unsupported. Mass alone cannot determine purity. An appropriate independent measurement would be needed, and the type of test depends on the substance. Transfer question: What single new observation, controlled comparison or source check would most improve confidence here? Record an answer in plain words before turning to the next case.
Decision 14: Unequal starting volumes
Proposed statement: One trial uses 100 g water and another uses 150 g. Reasoned response: Raw mass differences may reflect solvent amount rather than procedure effectiveness. Use relevant normalisation and only compare quantities whose bases and conditions are clear. Transfer question: What single new observation, controlled comparison or source check would most improve confidence here? Record an answer in plain words before turning to the next case.
Decision 15: A rounded observation
Proposed statement: The data are recorded to 0.1 g, and a report gives 16.43729 g. Reasoned response: The extra digits are not supported by the stated reading. Report measured precision faithfully and use unrounded internal calculations only when there are justified intermediate figures. Transfer question: What single new observation, controlled comparison or source check would most improve confidence here? Record an answer in plain words before turning to the next case.
Decision 16: A graph trend
Proposed statement: A solubility plot slopes upward with temperature. Reasoned response: The graph can support a relationship over its displayed range. It does not on its own establish the recovered mass from a particular physical experiment. Connect the model to starting conditions. Transfer question: What single new observation, controlled comparison or source check would most improve confidence here? Record an answer in plain words before turning to the next case.
Decision 17: A declaration of cause
Proposed statement: A student blames the filter for all shortfall. Reasoned response: Possible transfer residue is one explanation, not the only one. Plan observations that distinguish transfer loss from incomplete crystallisation and sample treatment. Transfer question: What single new observation, controlled comparison or source check would most improve confidence here? Record an answer in plain words before turning to the next case.
Decision 18: An uncertain sample
Proposed statement: One group describes its sample as damp. Reasoned response: That note should be preserved, not erased. It casts doubt on a dry-mass interpretation and suggests a controlled supervised mass check. Transfer question: What single new observation, controlled comparison or source check would most improve confidence here? Record an answer in plain words before turning to the next case.
Decision 19: A good public caption
Proposed statement: The label reads ‘Measured sample mass, not verified purity’. Reasoned response: Appropriate: it communicates the evidence and boundary of the claim. A reader can see what was observed and what remains unknown. Transfer question: What single new observation, controlled comparison or source check would most improve confidence here? Record an answer in plain words before turning to the next case.
Decision 20: A proposed test
Proposed statement: The group repeats the same wet weighing five times. Reasoned response: This may help assess reading consistency but does not directly settle dry mass or purity. Choose a test that answers the actual uncertainty, rather than counting repetitions. Transfer question: What single new observation, controlled comparison or source check would most improve confidence here? Record an answer in plain words before turning to the next case.
Decision 21: An average result
Proposed statement: The mean of 16.4, 17.8 and 15.9 is 16.7 g. Reasoned response: Numerically correct. Interpret with the range of 1.9 g and the variable procedures. Do not present 16.7 g as a proven population value from three inconsistent trials. Transfer question: What single new observation, controlled comparison or source check would most improve confidence here? Record an answer in plain words before turning to the next case.
Decision 22: A missing reference
Proposed statement: The final report says ‘almost the maximum’ without naming the model. Reasoned response: Specify the theoretical comparison, the recovered mass and the assumptions. A reader should not have to infer which maximum is intended. Transfer question: What single new observation, controlled comparison or source check would most improve confidence here? Record an answer in plain words before turning to the next case.
Decision 23: A report with no recommendation
Proposed statement: The findings are careful but the report ends abruptly. Reasoned response: If the task asks for action, add a proportional recommendation that addresses evidence gaps. Scientific honesty still needs a decision to be useful to its audience. Transfer question: What single new observation, controlled comparison or source check would most improve confidence here? Record an answer in plain words before turning to the next case.
Decision 24: An ambitious new claim
Proposed statement: The club wants to state both recovery and purity at an exhibition. Reasoned response: Treat them as separate claims with separate evidence. Publish recovery if its basis is clear; do not publish a purity percentage without appropriate analysis and documented verification. Transfer question: What single new observation, controlled comparison or source check would most improve confidence here? Record an answer in plain words before turning to the next case.
Longer worked case: the display board must be approved
A teacher receives a draft claiming: ‘The club found a reliable way to make almost perfectly pure crystals. Our best collection was 17.8 g from an ideal 20.0 g. We have proved the process works every time.’ Rewrite it in three movements. First state what is established: the group recorded 17.8 g in one trial, which is nominally 89% of the assumed maximum. Second state what is unresolved: sample drying, purity testing and consistency across comparable trials. Third recommend a controlled repeat and a separate check before making a purity claim.
One possible replacement reads: ‘Our group obtained crystals and recorded 17.8 g in the highest-mass trial. This is 89% of the reference maximum under the stated model. The trial does not verify chemical purity, and differences between collection conditions limit comparisons. We recommend consistent supervised collection and an appropriate verification method before using stronger labels.’ The replacement is shorter than the original but contains more usable knowledge.
A harder case: two explanations compete
Suppose the measured mass is lower than expected. Hypothesis A says material remained dissolved at the final temperature. Hypothesis B says the crystals formed but some were lost during transfer. Which data distinguish the hypotheses? A final-temperature record and an appropriate solubility comparison inform A. Observations or controlled measures of residue along the transfer path inform B. A lower recovered mass alone cannot choose between them.
An advanced answer states what finding would make each explanation more credible. If final conditions predict substantial dissolved material, A becomes plausible. If a carefully controlled transfer increases collected dry mass while cooling conditions remain comparable, B gains support. The aim is to separate causes with evidence rather than simply select the explanation that sounds most familiar.
A final rehearsal for a short oral explanation
Imagine the teacher asks, ‘Was this experiment a success?’ A clear answer is: ‘It demonstrated that crystals can form when the solution cools. The recorded mass was below the model maximum, but the collection method and drying conditions introduce uncertainty. We cannot infer purity from recovery mass. We should repeat the trial under controlled conditions and use a suitable purity check if we want to make a composition claim.’
That answer is not a script to memorise for English Paper 4, because Paper 4 has its own video-based planned response and spoken interaction format. It is practice in organising an evidence-based explanation. A good oral response is intelligible, accurate and willing to separate known facts from decisions still to be made.
Questions families and students often ask
Is an 82% recovery a poor mark? It is a scientific ratio in this exercise, not a student examination score. Its quality depends on the experiment’s aim and conditions. Are crystals necessarily pure? No. Their appearance and presence are not a certified composition test. Can three repeats prove reliability? They can provide more information, but shared biases and inconsistent methods still matter.
Do combined-science students follow K324? No; consult their registered combination and the applicable K326, K327 or K328 syllabus. Why include English in science work? Because unsupported wording can turn a correct calculation into a false public claim. What should I revise next? Study a different separation method, then practise interpreting its measurements and limits.
Connected reading and official specifications
Use the SEAB SEC qualification overview, K300 English, K310 Mathematics and <a href="https://isomer-user-content.by.gov.sg/334/db643e6e-c6a2-4da8-9a29-f12b8cbadcc9/K324 Chemistry for examination specifications. Combined-science learners should read the K326/K327/K328 syllabus. Sengkang’s PSLE-to-G3 transition guide and G3 Science owner connect this case to the wider study route.
