How to perform in the new G2 SEC Science examination when a structured question looks dense is to decompose it before answering. A long prompt may contain context, data, a diagram, several conditions, an unfamiliar object and a command word in one block. The learner who tries to process all of it at once can feel that the Science is harder than it really is.
This thirty-second Learner’s Guide provides a decomposition system: task → evidence → scientific operation → answer. The purpose is not to reduce every Science question to a template. It is to separate the information so the learner can see which part of the prompt tells the story, which part provides evidence, which scientific idea must be used and what the final response must actually do.
For 2027, SEAB lists G2 Science as K223 Science (Physics, Chemistry), K224 Science (Physics, Biology) and K225 Science (Chemistry, Biology). The shared G2 Science assessment objectives include locating, selecting, organising and presenting information; translating information between forms; manipulating data; identifying patterns; drawing inferences; providing reasoned explanations; making predictions; and solving unfamiliar problems using syllabus principles. Use the SEAB 2027 G2 syllabus page for current official details.
Long Questions Are Often Several Small Jobs
A long Science question may feel difficult because its parts are visually compressed. One sentence introduces an unfamiliar device. A table gives measurements. A diagram labels parts. The final line asks for an explanation.
The learner should separate these jobs before deciding whether the underlying Science is actually difficult.
The Four-Part Decomposition
1. Task
What does the command word require: state, describe, explain, predict, suggest, calculate, determine, justify or evaluate? The command decides the answer form.
2. Evidence
What information in the question must be used? This may be an observation, value, graph trend, diagram feature, experimental condition or comparison.
3. Scientific operation
What thinking turns the evidence into the answer? Recall a concept, calculate a quantity, compare values, infer a relationship, apply a mechanism, evaluate a method or transfer a known principle to an unfamiliar setting.
4. Answer
What is the smallest complete response that performs the task using the evidence and operation?
Keeping these parts separate prevents the learner from writing topic knowledge before understanding the job.
The Question Strip
During practice, draw four short labels beside a structured question:
- T = task;
- E = evidence;
- O = operation;
- A = answer.
The labels can disappear later when the process becomes internal. Their purpose is to make the structure visible while the skill is being built.
Task: Find the Verb First
The command word is often the fastest way to reduce complexity. A page of unfamiliar context may still end with a simple “calculate”, “describe” or “explain”.
Read the verb before deciding how difficult the question is.
Evidence: Separate Given From Known
Science answers combine two sources of information: what the question gives and what the learner knows from the syllabus. Keep them separate initially.
Given evidence may include a temperature change, colour, graph pattern or apparatus setup. Known science may include energy transfer, particle behaviour, diffusion, electrical relationships or biological processes.
The answer connects them.
Operation: Name the Thinking Move
Students often know the topic but do not know what to do with it. Naming the operation helps:
- compare;
- calculate;
- infer;
- apply concept;
- explain mechanism;
- evaluate method;
- predict from trend;
- translate graph to words;
- translate words to equation.
The operation is the bridge between evidence and response.
Answer: Complete the Job, Then Stop
A structured answer should be long enough to complete the task and no longer. If the question asks for one observation, do not write a mechanism. If it asks for an explanation, do not stop at the observation.
Decomposition protects response depth because the learner can see which job is actually required.
Context Is Not Always Tested Content
An unfamiliar context can include many details that are present only to make the scenario realistic. The learner should identify which details affect the scientific relationship.
Do not spend time learning the story during the examination. Extract the evidence relevant to the task.
The Context–Evidence Separation
Use two questions: What is happening? and What information actually changes my answer?
The first gives orientation. The second identifies evidence.
Unfamiliar Names
A question may introduce a device, organism or material with an unfamiliar name. Do not assume the name itself is knowledge you were expected to memorise.
Ask what role it plays in the question: source, conductor, reactant, sample, sensor, organ, barrier, input, output or measured object. Role often matters more than name.
The Role-Label Method
Replace the unfamiliar noun temporarily with its role: “device A” becomes “temperature sensor”, “material B” becomes “insulator”, “organism C” becomes “sample organism”.
This reduces novelty and exposes the known Science inside the question.
Data Is Evidence, Not Decoration
If a table or graph is provided, assume it has a job. Identify variables, units, pattern and any important exception before explaining.
Do not write a general science explanation that ignores the data supplied.
The Data Strip
- Name the independent variable.
- Name the dependent variable.
- Read units and scale.
- Describe the pattern.
- Identify anomaly or boundary if relevant.
- Then apply the scientific concept.
This sequence separates description from explanation.
Diagrams Are Information Structures
A diagram may show arrangement, connection, direction, location or apparatus. Read labels and relationships before deciding what concept is being tested.
Do not interpret a diagram by appearance alone. Use the information the diagram explicitly provides.
The Diagram Strip
- What are the labelled parts?
- What is connected to what?
- What direction or pathway is shown?
- Which quantities or conditions are attached?
- What feature changes between cases?
Once the diagram is reduced to relationships, the scientific operation becomes easier to identify.
Multi-Part Questions Need a Reset at Every Subpart
Part (b) may use information from part (a), but it may also introduce a new command word or new evidence. Do not carry the answer form automatically from one subpart to the next.
Read each subpart as a new T-E-O-A sequence.
Structured Physics Questions
Physics structured questions often combine a diagram, values, units and several subparts. Decompose them by quantities first.
- Task: calculate, describe, explain, compare or predict?
- Evidence: values, graph features, diagram labels, direction or units.
- Operation: select relationship, calculate, infer physical behaviour or explain mechanism.
- Answer: numerical value with unit, comparison or physical explanation.
If the context is unfamiliar, the quantities often reveal the familiar Physics underneath.
Physics Example Pattern
A question introduces an unfamiliar lifting device, gives mass, time and height, then asks why one design is more effective. Do not study the machine name. Identify the quantities, calculate the relevant physical measure if requested, then compare using the result.
The device is context. The quantities are evidence. The physical relationship is the operation.
Structured Chemistry Questions
Chemistry structured questions often mix observation, apparatus, substances, data and particle reasoning. Keep observation and explanation separate.
- Task: state observation, identify substance, explain reaction, calculate, suggest or evaluate?
- Evidence: colour, gas, precipitate, temperature, mass, volume, pH or reaction conditions.
- Operation: apply particle/reaction model, calculate, compare or infer.
- Answer: precise chemical statement at the requested level.
The question may contain several true chemical facts, but only some are relevant to the command.
Chemistry Example Pattern
A question gives two substances, a temperature table and an unfamiliar industrial context. If the task asks for the trend, describe the data first. If the next subpart asks why, move to particle or reaction explanation. Do not answer both jobs in the first subpart.
Structured Biology Questions
Biology questions often contain longer causal chains. Decompose them by structure, process and outcome.
- Task: describe, explain, predict, compare or suggest?
- Evidence: organ structure, graph, condition, measured response or experimental result.
- Operation: connect structure to function, condition to process or process to outcome.
- Answer: causal biological chain with the requested depth.
The learner should avoid dumping all known facts about the system. Follow the evidence supplied in this question.
Biology Example Pattern
An unfamiliar plant species is shown with a graph of gas exchange under different conditions. The species name is context. The graph is evidence. The relevant biological process is the operation. The answer should connect the pattern to the process rather than describe the plant generally.
Combined Science: Decompose Before Switching Modes
For K223, K224 and K225, the learner also needs discipline orientation. Add one label before T-E-O-A: Physics, Chemistry or Biology.
The full sequence becomes discipline → task → evidence → operation → answer. This prevents a learner from carrying the wrong scientific model from the previous question.
The Five-Box Method
- Discipline.
- Command word.
- Evidence.
- Scientific operation.
- Answer form.
During practice, draw five small boxes. In the real paper, the boxes can become a mental sequence.
Structured Questions With Calculations
Calculations often sit inside a larger reasoning question. Separate the numerical job from the interpretive job.
If the question asks “calculate and hence explain”, the calculation is evidence for the later explanation. Do not stop at the number. Conversely, do not explain before the required value has been obtained.
Calculation Decomposition
- Target quantity;
- known quantities and units;
- relationship or formula;
- substitution;
- result;
- interpretation if required.
This protects the learner from calculator-first behaviour.
Structured Questions With Graphs
A graph question may contain several jobs: read, calculate gradient, describe trend, explain pattern and predict beyond a point. Treat each command separately.
The same graph is evidence for multiple operations. The operation changes with the subpart.
Structured Questions With Tables
Tables can contain far more values than one subpart needs. Read the row or column labels, identify the variables and select only the data relevant to the command.
Do not copy large sections of the table into the answer. Use enough values to support the statement.
Structured Questions With Experiments
Experimental questions often combine method, data and evaluation. The decomposition becomes:
- Aim or relationship being tested;
- variables and apparatus;
- data or observation;
- limitation or source of error;
- requested operation: explain, evaluate, suggest improvement or conclude.
Use Vol 0020 for the deeper experimental-question system.
Structured Questions With Unfamiliar Information
The shared G2 Science assessment objectives allow questions based on information unfamiliar to the candidate, provided the required principles are within the syllabus. This is why decomposition matters. Novel context does not automatically mean novel Science.
Strip unfamiliar wording down to evidence and role. Then use the known principle logically.
The Novelty Filter
When a strange term appears, ask:
- Do I need to know this name, or does the question define its role?
- What measurable or observable information is attached to it?
- Which known concept connects to that information?
- What exactly does the command ask me to do with it?
The novelty filter prevents unfamiliar nouns from consuming attention.
Ignore Decorative Detail
Some contextual details create realism but do not affect the scientific answer. During practice, cross out information that can be removed without changing the question’s solution.
This teaches relevance. Do not cross out information in the real paper unless you are sure it is irrelevant; the mental skill is to distinguish evidence from narrative.
The Evidence Hierarchy
Not all information has equal weight. Direct measurements and stated conditions usually outrank assumptions. A graph trend outranks a vague expectation. An observed colour change is evidence; a guessed reason is not.
Build answers from the strongest evidence available.
The Scientific Operation Menu
- recall a definition;
- apply a concept;
- calculate;
- compare;
- describe trend;
- infer;
- explain mechanism;
- predict;
- suggest;
- evaluate;
- conclude.
When stuck, name the operation. The answer often becomes clearer once the thinking job is explicit.
Operation Versus Content
A learner may know the content but choose the wrong operation. For example, they know reaction-rate factors but the question asks only to describe the graph. Or they know photosynthesis but the question asks to compare two data sets.
Separate “what Science do I know?” from “what must I do with it?”
The Answer-Form Menu
- single fact or term;
- data-supported description;
- calculation with unit;
- causal explanation;
- comparison statement;
- prediction plus basis;
- suggestion plus reason;
- evaluation of method or evidence;
- bounded conclusion.
The answer form should match the command word, not the amount of Science the learner remembers.
The First-Line Rule for Structured Science
After decomposition, the learner should be able to write one useful first line. It might be a trend statement, a formula, a named concept, an observation or a causal starting point.
If no valid first line is available, return to task and evidence. The problem may still be in interpretation rather than Science knowledge.
The One-Evidence Rule
Before writing an explanation or conclusion, identify at least one piece of evidence from the question that the answer must account for.
This prevents generic topic answers that could have been written without reading the prompt.
The One-Mechanism Rule
For “explain” questions, identify the mechanism before adding detail. One correct mechanism developed clearly is usually more valuable than several unrelated keywords.
The mechanism should connect the evidence to the outcome.
The One-Boundary Rule
For predictions, suggestions and conclusions, identify the boundary of what the evidence supports. The answer should not claim more than the data, context or model allows.
Scientific restraint is part of structured-question control.
The Structured-Question Error Ledger
- command word missed;
- context mistaken for tested content;
- relevant data ignored;
- irrelevant detail copied into answer;
- operation not identified;
- description written instead of explanation;
- calculation completed but interpretation omitted;
- wrong discipline model selected;
- conclusion broader than evidence;
- one subpart’s answer form carried into the next subpart.
These errors are different from content gaps. A student can know the topic and still fail the question structure.
The Decomposition-Only Drill
Take ten structured questions and do not answer them. For each, label discipline, task, evidence, operation and expected answer form.
This isolates question reading from content production and makes interpretation weaknesses visible.
The Evidence-Selection Drill
Give the learner a long prompt with more information than the answer needs. Ask them to underline only the evidence relevant to one subpart and explain why the rest is not needed.
This trains relevance without requiring a full written response.
The Operation-Choice Drill
Present a question and several possible operations: calculate, compare, explain, infer or evaluate. Choose the correct operation before using any subject knowledge.
The learner begins to recognise that knowing the chapter is not the same as knowing the task.
The Answer-Form Drill
Provide a command word and evidence, then ask the learner to state the expected form before answering: one term, one calculation, one comparison sentence, a causal chain or an evaluation.
This reduces over-answering and under-answering.
The Mixed-Subpart Drill
Use one structured question with four subparts using different commands. Require a visible reset before each subpart.
The goal is to stop answer-form carryover: a long explanation after a “state” command, or a bare observation after an “explain” command.
The Unfamiliar-Context Drill
Use questions with unfamiliar devices, organisms or materials but familiar syllabus principles. The learner must replace the unfamiliar name with its functional role before answering.
This trains transfer and reduces the intimidation caused by novelty.
The Data-to-Explanation Drill
- Describe the graph or table without explanation.
- Identify the relevant scientific concept.
- Explain the mechanism separately.
- Combine only after both parts are correct.
This teaches the learner not to blur description and explanation.
The Calculation-to-Interpretation Drill
After every selected Science calculation, add one sentence explaining what the value means in the context. Then decide whether the original question actually required that interpretation.
The drill develops flexibility: calculate when asked to calculate; interpret when asked to interpret.
The Method-Evaluation Drill
Give a short experimental method. The learner identifies aim, variables, measurement, one limitation, its consequence and one targeted improvement.
This is a T-E-O-A sequence: evaluate is the task, method details are evidence, judgement is the operation and the improvement is part of the answer.
The 20-Minute Structured-Question Session
- Five minutes: decompose four questions without answering.
- Ten minutes: answer two of them fully.
- Five minutes: compare answer form with command word and evidence.
Short decomposition sessions can be used frequently without waiting for a full Science paper.
A Four-Week Structured-Question Build
Week 1 — task and evidence
Focus on command words, variables, graph/table reading and relevant information selection.
Week 2 — operation
Practise deciding whether the question needs calculation, comparison, inference, mechanism or evaluation.
Week 3 — unfamiliar contexts
Use novel scenarios while keeping syllabus principles familiar. Train functional role labels.
Week 4 — timed transfer
Complete realistic structured sections. Record where decomposition becomes rushed and which type of error appears first.
Decomposition Under Time
The real examination does not allow a long annotation ritual. The practice labels should compress into a rapid mental sequence:
- What job?
- What evidence?
- What Science move?
- What final form?
Four questions, a few seconds.
Decomposition Under Fatigue
Late in the paper, learners are more likely to answer from topic familiarity rather than question structure. Use the command word as the reset anchor.
If the answer begins before the learner can name the job, slow down for one sentence. That small pause can protect several marks.
Decomposition After a Mistake
If a later check shows the answer is wrong, revisit the four parts. Was the task misread? Was evidence misselected? Was the operation invalid? Or was the answer expression incomplete?
This locates the error more efficiently than rewriting the entire response.
Decomposition and Command Words
Use Vol 0024 for deeper command-word control. This volume places that command inside a larger question-reading system.
Decomposition and Unfamiliar Contexts
Use Vol 0016 for transfer. Decomposition is the method that separates unfamiliar surface context from familiar syllabus principle.
Decomposition and Combined-Science Switching
Use Vol 0028. Add discipline orientation before task-evidence-operation-answer so the correct Physics, Chemistry or Biology model is activated.
Decomposition and Mark Security
Use Vol 0025. Many accessible Science marks are lost because a dense question is treated as difficult before it is decomposed.
Use Examination Craft for Whole-Paper Control
For pacing, returns and checking, continue through the Examination Craft hub. Decomposition should make questions faster to enter, not become another time-consuming procedure.
The PSLE Bridge
The PSLE rule Evidence Before Explanation remains the foundation. G2 structured-question decomposition adds task and operation around that evidence.
Final Rule
A long Science question is not one giant problem. It is a sequence of small decisions.
Name the discipline. Read the command. Select the evidence. Identify the scientific operation. Produce the smallest complete answer. Then reset for the next subpart. When the structure becomes visible, unfamiliar questions become much easier to control.
Applied Structured-Question Scenarios
A Physics question may describe an unfamiliar transport system, give distances and times and finally ask for average speed. The story looks technical, but the task is calculate. Evidence is distance and time. Operation is the speed relationship. Answer is a value with unit.
A later subpart may ask why one design requires less force. The task has changed to explain. The earlier numerical result may become evidence, but the operation is now physical reasoning.
Chemistry Scenario
A table shows temperature against time for two reactions and the question asks which reaction is faster. Evidence is the rate at which the measured change occurs. Operation is comparison. A later “why” subpart changes the operation to a chemical mechanism.
The same data can support several different jobs; the command word decides which job is current.
Biology Scenario
An unfamiliar plant is shown with a graph of gas exchange under different conditions. The plant name is context. The graph is evidence. The biological process is the operation. The answer connects the pattern to the process rather than describing the organism generally.
Experimental Scenario
A method changes concentration, measures time and controls temperature. One subpart asks for the independent variable; another asks why temperature is controlled; another asks for an improvement. These are identify, explain and evaluate. The learner should reset at every subpart.
The Two-Layer Evidence Model
Structured Science often contains direct evidence and derived evidence. Direct evidence is given: a reading, observation, graph or condition. Derived evidence is produced by the learner: a calculated value, comparison or inferred pattern.
Later subparts may depend on the derived evidence. Label it clearly so the reasoning chain remains visible.
The Evidence-to-Mechanism Chain
- evidence or condition;
- relevant concept;
- mechanism;
- outcome.
If one link is missing, the explanation may become a jump rather than a causal chain.
The Evidence-to-Judgement Chain
- claim or choice;
- evidence;
- criterion or scientific reason;
- judgement.
This prevents evaluation from becoming a list of observations with no conclusion.
The Evidence-to-Prediction Chain
- observed trend or known principle;
- changed condition;
- expected outcome;
- reason if requested.
A prediction is a structured extension of evidence, not a guess.
The Evidence Boundary
Every structured answer should know where the evidence stops. If an experiment tested a limited range, do not automatically claim the same relationship outside it. If the sample is small, avoid unrestricted population claims.
The boundary protects scientific validity.
The Hidden-Assumption Audit
During practice, underline assumptions your answer makes that the question did not state. Some assumptions are legitimate within the syllabus model; others are inventions.
If the answer depends on an invented assumption, look again for evidence or choose a more cautious conclusion.
Structured-Question Triage
When time is short, prioritise subparts with clear task-evidence-operation routes. Mark blocked subparts for return.
A dense question may still contain several secure marks. Do not abandon the whole question because one subpart is difficult.
The Subpart Independence Check
Read later subparts even if an earlier one is blocked. A later question may use new information or ask a separate concept.
The question structure determines dependency. Do not create dependency mentally where none exists.
The 60-Second Decomposition Practice
Give the learner a long question and exactly sixty seconds to identify discipline, command word, three pieces of relevant evidence, one scientific operation and the expected answer form.
Do not answer yet. This builds speed in question entry.
The 5-Minute Structured-Answer Practice
After decomposition, give five minutes to answer. Then compare the response with the original plan. Did the answer stay aligned to the task and evidence?
This exposes drift between interpretation and execution.
The Reverse-Decomposition Review
- What answer did I actually give?
- What operation does that answer show?
- What evidence did I use?
- What task did I think I was answering?
- Does that match the real question?
This is useful when a scientifically correct-looking answer still loses marks.
The Structured-Question Notebook
Keep a compact log of wrong question structures, not only wrong topics: ignored table evidence, answered describe as explain, missed discipline switch, calculated but did not interpret.
Patterns in this notebook reveal performance weaknesses that chapter-based revision can miss.
Decomposition Under Fatigue
Late in the paper, learners are more likely to answer from topic familiarity rather than question structure. Use the command word as the reset anchor.
If the answer begins before the learner can name the job, slow down for one sentence. That pause can protect several marks.
Decomposition After a Mistake
If a check shows the answer is wrong, revisit the four parts. Was the task misread? Was evidence misselected? Was the operation invalid? Or was the answer expression incomplete?
This locates the error more efficiently than rewriting the entire response.
Use the Existing G2 Science System
Use Vol 0024 for command words, Vol 0016 for unfamiliar contexts, Vol 0028 for discipline switching and Vol 0025 for mark security.
Use Examination Craft
For pacing, returns and checking, continue through the Examination Craft hub. Decomposition should make questions faster to enter, not become another time-consuming procedure.
The PSLE Bridge
The PSLE rule Evidence Before Explanation remains the foundation. G2 structured-question decomposition adds task and operation around that evidence.
Final Rule
A long Science question is not one giant problem. It is a sequence of small decisions.
Name the discipline. Read the command. Select the evidence. Identify the scientific operation. Produce the smallest complete answer. Then reset for the next subpart. Once the structure becomes visible, unfamiliar questions become much easier to control.
The Structured-Question Finish
The final refinement is to make decomposition fast enough that it disappears into normal reading. The learner should not need to draw boxes on every real examination question. The practice structure becomes four internal questions: what is the job, what is the evidence, what scientific move connects them, and what form must the answer take?
This sequence is especially useful when the question feels intimidating because of length. Long prompts often become ordinary once the context is separated from the scientific operation.
The Evidence-Selection Finish
Before writing, identify which pieces of information the examiner deliberately supplied for this subpart. A graph, diagram or table may contain more data than required. Select enough evidence to support the answer without copying the whole source.
Evidence selection is itself a scientific skill. It shows that the learner can distinguish signal from context.
The Operation Finish
If the learner knows the topic but remains stuck, name the operation out loud during practice: compare, calculate, infer, explain, predict, evaluate. The operation often reveals the next step because it narrows the range of possible responses.
A learner who cannot name the operation may still be trying to answer from topic familiarity rather than question structure.
The Answer-Form Finish
- State/name → direct term or fact.
- Describe → evidence-based pattern or observation.
- Explain → causal mechanism.
- Calculate → relationship, working and unit.
- Compare → both sides on the same feature.
- Predict → expected outcome from trend or principle.
- Suggest → plausible evidence-based proposal.
- Evaluate → judgement with consequence and limitation.
- Conclude → bounded claim supported by the evidence.
This final form check catches a class of mistakes that content review alone cannot detect.
The Advanced Structured-Question Standard
An advanced learner can enter an unfamiliar question without waiting for recognition of the exact example. They orient to the discipline, read the command, select the relevant evidence and choose a scientific operation. Novelty remains on the surface while the reasoning stays familiar.
That is the deeper purpose of decomposition. It does not make the question simpler than it is; it makes the structure visible enough for the learner’s actual Science to operate.
Final Perspective
A structured question is a designed information environment. Some details establish context. Some details are evidence. The command word sets the job. The syllabus concept supplies the mechanism. The learner’s answer brings those parts together.
When that architecture becomes automatic, dense questions stop feeling like walls of text. They become sequences of manageable scientific decisions.
The Final Decomposition Rule
When a structured question still feels difficult after decomposition, the learner can identify the remaining bottleneck accurately. If the task and evidence are clear but no operation is available, the problem is likely a concept or transfer gap. If the operation is clear but the answer form is wrong, the problem is response control.
This diagnostic value matters beyond one paper. Decomposition turns “I cannot do this Science question” into a specific learning job that can be repaired.
The learner can then revise the precise weak link instead of doing another random worksheet. A decomposition habit therefore improves both examination performance and the quality of the revision that follows each practice paper.
