G3 Chinese writing needs developed, relevant ideas, but memorising whole model essays is a fragile way to produce them. A memorised paragraph may not answer the actual prompt, may carry the wrong tone or genre and may collapse when one line is forgotten. A stronger method stores portable reasoning moves: define the issue, take a position, explain a mechanism, choose evidence, consider a condition and reach a judgement. For narrative writing, the portable units are character goal, pressure, decision, consequence and change. Students can then build an original response that fits the question in front of them.
The honest quick answer for parents is that this is a decision task before it is a language-display task. The learner must notice what the text or prompt requires, select the useful material, and express or choose an answer without drifting into a familiar but irrelevant response.
This guide uses the 2027 Singapore-Cambridge Secondary Education Certificate G3 Chinese syllabus K320. It explains the assessment boundary, but it does not turn every lesson into examination drilling. Strong preparation alternates between close diagnosis, guided practice, independent transfer and timed stability.
eduKateSG · Chinese task explainer
Find the closest route
Read one chapter for the immediate question, or follow all five from definition to independent review.
1. Separate useful models from memorised scripts
2. Build portable idea-generation moves
3. Develop evidence and mechanism
4. Adapt ideas to genre, audience and prompt
Contents
Chapter 1: Separate useful models from memorised scripts
Official task and first diagnosis
Chapter 5: Retest originality on unfamiliar tasks
Progress checks, FAQs and next links
Begin by naming what the student must produce. A multiple-choice practical-text item, an open-ended comprehension answer, an email and a public forum response all use Chinese, but they demand different reader actions. When the task type is blurred, students fall back on general habits: translate everything, copy a nearby line, write a memorised opening or choose the longest answer.
Official 2027 boundary
| Feature | Reference |
|---|---|
| Level | G3 Chinese Language |
| Syllabus code | K320 |
| Writing priority | Relevant, developed and organised response |
| Argument unit | Claim, reason, mechanism, evidence and judgement |
| Narrative unit | Goal, pressure, decision, consequence and change |
| Model use | Study decisions and techniques rather than memorise full prose |
| Core outcome | Flexible original writing under a fresh prompt |
The table defines scope. It should prevent two common distortions: under-preparation because the component looks small, and over-preparation that lets one component consume the whole subject. Weighting informs balance; the learner’s actual error pattern informs the next lesson.
The first useful definition
G3 Chinese writing needs developed, relevant ideas, but memorising whole model essays is a fragile way to produce them. A memorised paragraph may not answer the actual prompt, may carry the wrong tone or genre and may collapse when one line is forgotten. A stronger method stores portable reasoning moves: define the issue, take a position, explain a mechanism, choose evidence, consider a condition and reach a judgement. For narrative writing, the portable units are character goal, pressure, decision, consequence and change. Students can then build an original response that fits the question in front of them.
Notice the verbs: identify, select, connect, respond and check. These operations make the task teachable. “Improve Chinese” is too broad for a twenty-minute practice. “Find the condition that changes the price” or “answer both parts with one supporting detail each” can be demonstrated and reviewed.
Six plausible failures
- 1. The student is choosing a prompt only because it resembles a memorised essay.
- 2. The student is forcing prepared examples into an unrelated question.
- 3. The student is copying sophisticated language without controlling its meaning.
- 4. The student is listing many points without explaining any mechanism.
- 5. The student is mistaking originality for inventing unrealistic content.
- 6. The student is rejecting all model texts instead of studying how good decisions work.
Each failure can produce an answer that looks busy. That is why marking only right or wrong is not enough. Ask what the student noticed first, where the evidence came from, and which check was used. A wrong answer reached through a sound process may need vocabulary repair; a lucky correct answer reached by guessing still needs method.
A calm baseline
Use one fresh item and do not teach before the first attempt. Ask the student to state the job, mark the controlling words and complete the answer. Record the time to start, the first evidence choice, any prompt required and the final check. Keep the original work. It will become the comparison point after instruction.
Avoid turning the baseline into a judgement. The purpose is to locate the first unstable decision. If the student understands the task but cannot retrieve Chinese, teach expression. If the student writes fluently but answers a different question, teach task reading and relevance. If the student succeeds untimed but collapses on screen or under a clock, teach control conditions.
Official sources
SEAB lists G3 Chinese under the 2027 SEC syllabus page. The detailed requirements appear in the K320 official syllabus PDF. Families should confirm the applicable year because assessment documents can change. School instructions remain important for day-to-day learning and examination administration.
Evidence is not merely a quotation. It is the smallest part of the source that justifies an answer or response decision. The student should be able to say, “I chose this because these words establish the audience, condition, cause, change or requested action.” That sentence makes guessing visible and gives feedback somewhere to attach.
Five evidence-finding demonstrations
Case 1: Prompt unpacking
Source: A familiar topic appears with a new demand word.
Move: Define exactly what must be explained, evaluated or narrated.
Reason: Topic similarity does not guarantee task similarity.
Teaching sequence: Let the student point to or state the evidence before producing the answer. Ask for one uninterrupted first attempt. Then compare the work with the task, not with a perfect model. Correct the earliest decision that caused the later error.
Transfer: Change the surface detail while preserving the underlying relationship. The student should complete the new version without copying the earlier wording. Delayed success on changed material is stronger evidence than a neat immediate rewrite.
Case 2: Position
Source: An argument begins with background but no answer.
Move: State a defensible judgement that can guide paragraph choices.
Reason: A position filters ideas.
Teaching sequence: Let the student point to or state the evidence before producing the answer. Ask for one uninterrupted first attempt. Then compare the work with the task, not with a perfect model. Correct the earliest decision that caused the later error.
Transfer: Change the surface detail while preserving the underlying relationship. The student should complete the new version without copying the earlier wording. Delayed success on changed material is stronger evidence than a neat immediate rewrite.
Case 3: Reason generation
Source: The student can think of only one obvious point.
Move: Ask who is affected, what changes, why and under what condition.
Reason: Structured questions generate distinct reasons.
Teaching sequence: Let the student point to or state the evidence before producing the answer. Ask for one uninterrupted first attempt. Then compare the work with the task, not with a perfect model. Correct the earliest decision that caused the later error.
Transfer: Change the surface detail while preserving the underlying relationship. The student should complete the new version without copying the earlier wording. Delayed success on changed material is stronger evidence than a neat immediate rewrite.
Case 4: Mechanism
Source: A paragraph says an action is beneficial.
Move: Explain how it changes access, behaviour, attention or outcome.
Reason: Mechanism turns assertion into development.
Teaching sequence: Let the student point to or state the evidence before producing the answer. Ask for one uninterrupted first attempt. Then compare the work with the task, not with a perfect model. Correct the earliest decision that caused the later error.
Transfer: Change the surface detail while preserving the underlying relationship. The student should complete the new version without copying the earlier wording. Delayed success on changed material is stronger evidence than a neat immediate rewrite.
Case 5: Evidence
Source: A memorised example only loosely fits.
Move: Choose a credible scenario that directly demonstrates the mechanism.
Reason: Evidence serves the point.
Teaching sequence: Let the student point to or state the evidence before producing the answer. Ask for one uninterrupted first attempt. Then compare the work with the task, not with a perfect model. Correct the earliest decision that caused the later error.
Transfer: Change the surface detail while preserving the underlying relationship. The student should complete the new version without copying the earlier wording. Delayed success on changed material is stronger evidence than a neat immediate rewrite.
Question words create search instructions
Teach question language as an action system. 谁 asks for a person or group. 哪里 asks for place. 什么时候 asks for time. 为什么 asks for cause or purpose. 怎样 may ask for method, process or manner. 哪一点 or 哪两点 sets the amount. 有什么作用 asks for function, not repetition. The exact wording determines what kind of evidence is useful.
Students should not circle every keyword until the page becomes noise. Mark the command, required number of points and any boundary such as “according to the passage”, “before”, “except” or “most suitable”. Then search with a purpose. The aim is less marking and more control.
Evidence before expression
When written Chinese is difficult, separate selection from expression. Let the student first identify the correct source line, image detail or prompt point. If selection is correct, build the shortest complete Chinese response. Then improve clarity. This order protects genuine comprehension from being mistaken for a total language failure.
Conversely, fluent expression should not hide weak selection. Ask the student to trace each major claim back to the source or prompt. If no relevant evidence exists, the sentence is probably decoration, assumption or drift.
A response should carry the task with the least unnecessary load. That does not mean every answer must be short. It means each sentence has a job: identify, explain, support, compare, request, recommend, qualify or close. When students can name paragraph and sentence jobs, length becomes meaningful instead of padded.
Five response-building demonstrations
Case 6: Condition
Source: A proposal is treated as universally effective.
Move: Identify the resources, participation or design needed for it to work.
Reason: Conditions create nuance.
Teaching sequence: Let the student point to or state the evidence before producing the answer. Ask for one uninterrupted first attempt. Then compare the work with the task, not with a perfect model. Correct the earliest decision that caused the later error.
Transfer: Change the surface detail while preserving the underlying relationship. The student should complete the new version without copying the earlier wording. Delayed success on changed material is stronger evidence than a neat immediate rewrite.
Case 7: Counterpoint
Source: The essay repeats support for one side.
Move: Consider a real limitation and respond to it.
Reason: A counterpoint can sharpen judgement.
Teaching sequence: Let the student point to or state the evidence before producing the answer. Ask for one uninterrupted first attempt. Then compare the work with the task, not with a perfect model. Correct the earliest decision that caused the later error.
Transfer: Change the surface detail while preserving the underlying relationship. The student should complete the new version without copying the earlier wording. Delayed success on changed material is stronger evidence than a neat immediate rewrite.
Case 8: Narrative goal
Source: A story begins with long description.
Move: Give the character a clear want or responsibility.
Reason: Goal creates direction.
Teaching sequence: Let the student point to or state the evidence before producing the answer. Ask for one uninterrupted first attempt. Then compare the work with the task, not with a perfect model. Correct the earliest decision that caused the later error.
Transfer: Change the surface detail while preserving the underlying relationship. The student should complete the new version without copying the earlier wording. Delayed success on changed material is stronger evidence than a neat immediate rewrite.
Case 9: Narrative pressure
Source: Events occur but no meaningful problem develops.
Move: Introduce a constraint that forces a decision.
Reason: Pressure creates significance.
Teaching sequence: Let the student point to or state the evidence before producing the answer. Ask for one uninterrupted first attempt. Then compare the work with the task, not with a perfect model. Correct the earliest decision that caused the later error.
Transfer: Change the surface detail while preserving the underlying relationship. The student should complete the new version without copying the earlier wording. Delayed success on changed material is stronger evidence than a neat immediate rewrite.
Case 10: Narrative consequence
Source: The character makes a choice and the story ends.
Move: Show what changed externally or internally.
Reason: Consequence completes the arc.
Teaching sequence: Let the student point to or state the evidence before producing the answer. Ask for one uninterrupted first attempt. Then compare the work with the task, not with a perfect model. Correct the earliest decision that caused the later error.
Transfer: Change the surface detail while preserving the underlying relationship. The student should complete the new version without copying the earlier wording. Delayed success on changed material is stronger evidence than a neat immediate rewrite.
The complete-answer ladder
| Level | What the student produces | Next repair |
|---|---|---|
| 1. Fragment | A word or copied phrase | Add subject, relationship or requested detail |
| 2. Basic answer | The main idea only | Supply the missing part or evidence |
| 3. Supported answer | Idea plus relevant source or example | Explain the connection |
| 4. Controlled answer | Complete, relevant and appropriately expressed | Check precision, tone and efficiency |
| 5. Independent transfer | Same process on changed material | Increase time or language demand gradually |
The ladder is diagnostic rather than a permanent score. A student may be at different steps for different question types. Use it to decide the next prompt. Do not demand sophisticated language when the response is still missing its basic relationship.
Useful language grows around functions
Group phrases by their work. Cause language includes 因为、由于、因此 and 之所以…是因为…. Contrast includes 但是、然而、却 and 虽然…但是…. Requests need clear verbs and conditions. Recommendations need an action, reason and expected effect. Evidence links need phrases that show what an action or detail reveals.
Practise one function across several topics. The vocabulary then becomes portable. A memorised topic paragraph is brittle; a controlled cause explanation can support comprehension, email, forum writing and oral discussion.
Remove the scaffold on purpose
A scaffold is successful when it can disappear. Start with labelled fields or sentence prompts. Next remove the wording but keep the checklist. Then remove the checklist and ask the student to reconstruct it after the task. Finally, use an unfamiliar context under modest time. Record how much help was required at each stage.
Checking is a second reading with a narrower purpose. A vague instruction such as “check your work” often produces a quick glance. Give each pass a job. First check task coverage. Second check evidence and logic. Third check expression. Last check mechanics such as characters, punctuation, selection or submission.
Five control-and-checking demonstrations
Case 11: Model analysis
Source: A strong sample essay is provided.
Move: Label its prompt decisions, paragraph jobs and development moves.
Reason: Techniques transfer better than sentences.
Teaching sequence: Let the student point to or state the evidence before producing the answer. Ask for one uninterrupted first attempt. Then compare the work with the task, not with a perfect model. Correct the earliest decision that caused the later error.
Transfer: Change the surface detail while preserving the underlying relationship. The student should complete the new version without copying the earlier wording. Delayed success on changed material is stronger evidence than a neat immediate rewrite.
Case 12: Language borrowing
Source: One phrase from the model is useful.
Move: Adapt it only after checking meaning, collocation and register.
Reason: Selective learning avoids imitation.
Teaching sequence: Let the student point to or state the evidence before producing the answer. Ask for one uninterrupted first attempt. Then compare the work with the task, not with a perfect model. Correct the earliest decision that caused the later error.
Transfer: Change the surface detail while preserving the underlying relationship. The student should complete the new version without copying the earlier wording. Delayed success on changed material is stronger evidence than a neat immediate rewrite.
Case 13: Outline test
Source: The plan contains several attractive ideas.
Move: Match every paragraph to one prompt demand and distinct job.
Reason: Relevance should be visible before drafting.
Teaching sequence: Let the student point to or state the evidence before producing the answer. Ask for one uninterrupted first attempt. Then compare the work with the task, not with a perfect model. Correct the earliest decision that caused the later error.
Transfer: Change the surface detail while preserving the underlying relationship. The student should complete the new version without copying the earlier wording. Delayed success on changed material is stronger evidence than a neat immediate rewrite.
Case 14: Draft recovery
Source: A planned example stops fitting mid-paragraph.
Move: Return to the claim and replace the evidence rather than forcing it.
Reason: Structure supports recovery.
Teaching sequence: Let the student point to or state the evidence before producing the answer. Ask for one uninterrupted first attempt. Then compare the work with the task, not with a perfect model. Correct the earliest decision that caused the later error.
Transfer: Change the surface detail while preserving the underlying relationship. The student should complete the new version without copying the earlier wording. Delayed success on changed material is stronger evidence than a neat immediate rewrite.
Case 15: Fresh prompt
Source: The student practises only close variations.
Move: Change topic, stance or genre while reusing reasoning moves.
Reason: Variation tests flexibility.
Teaching sequence: Let the student point to or state the evidence before producing the answer. Ask for one uninterrupted first attempt. Then compare the work with the task, not with a perfect model. Correct the earliest decision that caused the later error.
Transfer: Change the surface detail while preserving the underlying relationship. The student should complete the new version without copying the earlier wording. Delayed success on changed material is stronger evidence than a neat immediate rewrite.
A repair log that predicts the next error
| Observed failure | Likely process gap | Next short drill |
|---|---|---|
| Missed a condition | Boundary words were not marked | Contrast ordinary rule with exception |
| Copied without answering | Question type was not named | Turn three commands into answer shapes |
| Incomplete prompt coverage | Points were not listed before drafting | Plan with one field per required point |
| Vague explanation | Evidence and link were merged | Write claim, evidence and because-link separately |
| Tone mismatch | Audience was not fixed | Rewrite one message for friend, teacher and public reader |
| Time collapse | No checkpoint protected later work | Rehearse with visible section deadlines |
Keep only recurring, high-value patterns. An error book containing every typo becomes too large to use. The next drill should be short enough to repeat and close enough to the original failure that improvement can be measured.
Time should be added after control
Untimed work reveals whether the method exists. Timed work reveals whether it is stable. If the learner cannot complete the process slowly, more speed practice usually automates guessing. Once the process is sound, shorten planning, increase source density or combine sections one variable at a time.
Use checkpoints instead of constant clock watching. Decide when the student should have selected a question, finished a section and begun review. A checkpoint is a navigation aid, not a threat. After practice, compare where time was spent with where marks and difficulty actually sit.
Feedback without taking over
Mark the first consequential error and ask a process question. “Which word limits who may join?” “What does why require?” “Which prompt point has no paragraph?” “Would this tone fit a public reader?” The student revises. Only supply final wording after a genuine attempt and explanation.
Finish with a fresh parallel item. The corrected original proves the teacher can improve a page. The parallel item shows whether the student learned. Keep both samples so progress conversations remain grounded in evidence.
A small weekly routine
- One ten-minute task-reading drill
- One evidence-selection set
- One complete response with guided checking
- One independent transfer task
- One delayed review of the recurring error
Adapt the amount to school work, CCA, rest and the student’s current stamina. Consistency beats a large plan that repeatedly fails to begin.
Progress is control becoming available sooner and with less rescue. Look for faster identification of the task, more relevant evidence, fewer missing parts, clearer links, appropriate tone and a check that catches an error before the teacher points it out. These signals usually appear before a dramatic examination result.
One complete self-checking model
Case 16: Final audit
Source: The essay sounds polished.
Move: Check prompt fit, independent development and whether any paragraph could belong to a different question.
Reason: Original relevance matters more than memorised fluency.
Teaching sequence: Let the student point to or state the evidence before producing the answer. Ask for one uninterrupted first attempt. Then compare the work with the task, not with a perfect model. Correct the earliest decision that caused the later error.
Transfer: Change the surface detail while preserving the underlying relationship. The student should complete the new version without copying the earlier wording. Delayed success on changed material is stronger evidence than a neat immediate rewrite.
Eight two-minute coaching cards
Card 1: Prompt unpacking
Read “A familiar topic appears with a new demand word.”. Before answering, ask whether you are choosing a prompt only because it resembles a memorised essay. Then define exactly what must be explained, evaluated or narrated. Topic similarity does not guarantee task similarity.
Complete a first response without interruption. Underline the evidence or prompt point that supports it. Cover any model and explain the decision aloud. On a later day, change one condition or audience and repeat. If the process survives the change with less prompting, record it as transfer rather than merely familiarity.
Card 2: Position
Read “An argument begins with background but no answer.”. Before answering, ask whether you are forcing prepared examples into an unrelated question. Then state a defensible judgement that can guide paragraph choices. A position filters ideas.
Complete a first response without interruption. Underline the evidence or prompt point that supports it. Cover any model and explain the decision aloud. On a later day, change one condition or audience and repeat. If the process survives the change with less prompting, record it as transfer rather than merely familiarity.
Card 3: Reason generation
Read “The student can think of only one obvious point.”. Before answering, ask whether you are copying sophisticated language without controlling its meaning. Then ask who is affected, what changes, why and under what condition. Structured questions generate distinct reasons.
Complete a first response without interruption. Underline the evidence or prompt point that supports it. Cover any model and explain the decision aloud. On a later day, change one condition or audience and repeat. If the process survives the change with less prompting, record it as transfer rather than merely familiarity.
Card 4: Mechanism
Read “A paragraph says an action is beneficial.”. Before answering, ask whether you are listing many points without explaining any mechanism. Then explain how it changes access, behaviour, attention or outcome. Mechanism turns assertion into development.
Complete a first response without interruption. Underline the evidence or prompt point that supports it. Cover any model and explain the decision aloud. On a later day, change one condition or audience and repeat. If the process survives the change with less prompting, record it as transfer rather than merely familiarity.
Card 5: Evidence
Read “A memorised example only loosely fits.”. Before answering, ask whether you are mistaking originality for inventing unrealistic content. Then choose a credible scenario that directly demonstrates the mechanism. Evidence serves the point.
Complete a first response without interruption. Underline the evidence or prompt point that supports it. Cover any model and explain the decision aloud. On a later day, change one condition or audience and repeat. If the process survives the change with less prompting, record it as transfer rather than merely familiarity.
Card 6: Condition
Read “A proposal is treated as universally effective.”. Before answering, ask whether you are rejecting all model texts instead of studying how good decisions work. Then identify the resources, participation or design needed for it to work. Conditions create nuance.
Complete a first response without interruption. Underline the evidence or prompt point that supports it. Cover any model and explain the decision aloud. On a later day, change one condition or audience and repeat. If the process survives the change with less prompting, record it as transfer rather than merely familiarity.
Card 7: Counterpoint
Read “The essay repeats support for one side.”. Before answering, ask whether you are choosing a prompt only because it resembles a memorised essay. Then consider a real limitation and respond to it. A counterpoint can sharpen judgement.
Complete a first response without interruption. Underline the evidence or prompt point that supports it. Cover any model and explain the decision aloud. On a later day, change one condition or audience and repeat. If the process survives the change with less prompting, record it as transfer rather than merely familiarity.
Card 8: Narrative goal
Read “A story begins with long description.”. Before answering, ask whether you are forcing prepared examples into an unrelated question. Then give the character a clear want or responsibility. Goal creates direction.
Complete a first response without interruption. Underline the evidence or prompt point that supports it. Cover any model and explain the decision aloud. On a later day, change one condition or audience and repeat. If the process survives the change with less prompting, record it as transfer rather than merely familiarity.
Six parent prompts that preserve student ownership
1. Ask about prompt unpacking
Show the learner “A familiar topic appears with a new demand word.” and ask, “What is this item making you decide?” Wait for the explanation before discussing vocabulary. Then ask the student to point to the exact source or prompt feature that controls the answer. Topic similarity does not guarantee task similarity.
After the response, do not immediately announce whether it is correct. Ask, “How would you check this without me?” The student should apply the relevant action: define exactly what must be explained, evaluated or narrated. Keep a note of the prompt required. At the next practice, use a changed example and offer less help. This makes the parent a calm observer of the process rather than the permanent supplier of answers.
2. Ask about position
Show the learner “An argument begins with background but no answer.” and ask, “What is this item making you decide?” Wait for the explanation before discussing vocabulary. Then ask the student to point to the exact source or prompt feature that controls the answer. A position filters ideas.
After the response, do not immediately announce whether it is correct. Ask, “How would you check this without me?” The student should apply the relevant action: state a defensible judgement that can guide paragraph choices. Keep a note of the prompt required. At the next practice, use a changed example and offer less help. This makes the parent a calm observer of the process rather than the permanent supplier of answers.
3. Ask about reason generation
Show the learner “The student can think of only one obvious point.” and ask, “What is this item making you decide?” Wait for the explanation before discussing vocabulary. Then ask the student to point to the exact source or prompt feature that controls the answer. Structured questions generate distinct reasons.
After the response, do not immediately announce whether it is correct. Ask, “How would you check this without me?” The student should apply the relevant action: ask who is affected, what changes, why and under what condition. Keep a note of the prompt required. At the next practice, use a changed example and offer less help. This makes the parent a calm observer of the process rather than the permanent supplier of answers.
4. Ask about mechanism
Show the learner “A paragraph says an action is beneficial.” and ask, “What is this item making you decide?” Wait for the explanation before discussing vocabulary. Then ask the student to point to the exact source or prompt feature that controls the answer. Mechanism turns assertion into development.
After the response, do not immediately announce whether it is correct. Ask, “How would you check this without me?” The student should apply the relevant action: explain how it changes access, behaviour, attention or outcome. Keep a note of the prompt required. At the next practice, use a changed example and offer less help. This makes the parent a calm observer of the process rather than the permanent supplier of answers.
5. Ask about evidence
Show the learner “A memorised example only loosely fits.” and ask, “What is this item making you decide?” Wait for the explanation before discussing vocabulary. Then ask the student to point to the exact source or prompt feature that controls the answer. Evidence serves the point.
After the response, do not immediately announce whether it is correct. Ask, “How would you check this without me?” The student should apply the relevant action: choose a credible scenario that directly demonstrates the mechanism. Keep a note of the prompt required. At the next practice, use a changed example and offer less help. This makes the parent a calm observer of the process rather than the permanent supplier of answers.
6. Ask about condition
Show the learner “A proposal is treated as universally effective.” and ask, “What is this item making you decide?” Wait for the explanation before discussing vocabulary. Then ask the student to point to the exact source or prompt feature that controls the answer. Conditions create nuance.
After the response, do not immediately announce whether it is correct. Ask, “How would you check this without me?” The student should apply the relevant action: identify the resources, participation or design needed for it to work. Keep a note of the prompt required. At the next practice, use a changed example and offer less help. This makes the parent a calm observer of the process rather than the permanent supplier of answers.
A monthly review table
| Question | Evidence |
|---|---|
| Does the student identify the task accurately? | One-sentence task description |
| Is evidence relevant and sufficient? | Marked source plus final answer |
| Is every required point present? | Prompt-to-response checklist |
| Does language carry the intended relationship? | Unedited first draft |
| Can the student detect a likely error? | Self-correction before feedback |
| Is the work sustainable? | Time, prompting and stress notes |
Compare like with like across several weeks. A harder passage or a more formal audience may temporarily lower accuracy even when the underlying process is improving. Note the task demand as well as the outcome.
Frequently asked questions
Are model essays useless?
No. Use them to study prompt interpretation, structure, development, vocabulary choices and revision decisions. Avoid memorising whole responses as universal scripts.
How can students generate more ideas?
Use structured questions about stakeholders, mechanism, cause, consequence, condition, comparison and limitation rather than waiting for inspiration.
Can examples be prepared in advance?
Students can build broad knowledge and flexible example families, but each example must be adapted to the actual claim and prompt.
How can teachers test whether an idea is transferable?
Change the topic, stance, audience or genre and ask the student to reuse the reasoning move without copying the original paragraph.
When outside support helps
Support is useful when the learner needs a repeated loop of diagnosis, modelling, attempt, feedback and transfer. Ask what the tutor observes during the first attempt, how corrections are prioritised, and how independence is measured. More worksheets are not automatically more teaching.
Small groups can provide explanation, peer language exposure and individual checking when students are actively producing work. eduKateSG’s small-groups tuition overview explains the setting. The actual fit should still be judged from the student’s evidence, school demands and weekly capacity.
Continue through the Chinese cluster
Read the main G3 Chinese guide for the wider subject map. Compare levels in What is G1, G2 and G3 Chinese?. Confirm current examination requirements on the official 2027 syllabus page and the K320 syllabus PDF.
The durable method is simple to state: define the job, locate the controlling evidence, build a complete response, check the relationship and test the method on changed material. It replaces guessing with something the learner can see, practise and own.
Related focused Chinese guides
G3 Chinese overview · G3 Chinese argumentative composition · G3 Chinese expository composition