Quick answer: PSLE Paper 2 comprehension asks students to build meaning at more than one depth. A useful architecture is literal → inferential → evaluative → evidence → answer. First establish what the text directly says; then connect clues to infer what is implied; evaluate ideas, language or viewpoints where the question requires it; identify the evidence that makes the interpretation defensible; and construct an answer that matches the exact question job.
SEAB’s 2026 English Language syllabus 0001 states that Paper 2 assesses candidates’ ability to use language appropriately in context and to comprehend written and multimodal texts at the literal, inferential and evaluative levels. These levels are connected. Weak literal understanding usually produces weak inference; weak evidence control makes evaluation drift into opinion.
Deep comprehension is not guessing what the writer “really means”. It is moving beyond the surface without leaving the evidence behind.
1. Literal Reading: Build the Shared Base
- who is involved;
- what happened;
- where and when;
- what a pronoun refers to;
- what a sentence states;
- what sequence the passage establishes.
Literal questions are not necessarily trivial. Dense syntax, unfamiliar vocabulary or multimodal information can make a directly stated idea difficult to extract.
2. Reference Chains Hold the Text Together
Track pronouns, repeated nouns and substitutions.
- Who does “he” refer to?
- What does “this decision” refer back to?
- Which earlier event is “the incident”?
- Does “they” refer to people, objects or organisations?
A wrong reference can distort an entire paragraph.
3. Connectors Reveal Relationships
| Signal | Relationship |
|---|---|
| however / but | contrast |
| because / since | cause |
| therefore / so | result |
| although / despite | concession |
| for example | illustration |
Students should not merely circle connectors. They should state what two ideas the connector relates.
4. Inference: Combine Clues Without Inventing
Text clue + relevant knowledge → bounded inference.
Good inference answers can usually point back to one or more details.
- What behaviour reveals emotion?
- What detail suggests motive?
- What consequence implies a cause?
- What comparison suggests the writer’s attitude?
If the answer cannot be linked to the text, it may be a plausible story rather than a supported inference.
5. Inference Strength Should Match Evidence Strength
| Evidence | Safer language |
|---|---|
| one weak clue | may / suggests |
| several consistent clues | likely / indicates |
| direct statement | states / shows directly |
6. Evaluation: Ask What Standard the Question Uses
Evaluative reading may require judgement about effectiveness, viewpoint, credibility, appropriateness or impact. The question supplies or implies the standard.
- effective for what purpose?
- convincing to which audience?
- credible based on what evidence?
- appropriate in which context?
- successful compared with what alternative?
7. Evaluation Is Not Personal Preference
“I like this advertisement” is not the same as explaining why it may persuade a target audience. A strong evaluation connects a feature to an effect and supports that connection with evidence.
Feature → intended effect → evidence → judgement.
8. Multimodal Reading: Words and Visuals Share the Meaning
- What does the image add that the words do not?
- Does the caption change how the image is read?
- What does layout emphasise?
- Which visual feature supports the main claim?
- Is information repeated, complemented or contrasted?
Do not treat a visual as decoration. It can carry evidence or change interpretation.
9. Vocabulary in Context
When meaning is unfamiliar, use the local sentence and surrounding paragraph.
- part of speech;
- contrast or example nearby;
- tone;
- cause/effect relationship;
- word family.
Then test the inferred meaning by substituting it back into the sentence.
10. Paragraph Function Helps Locate Answers
- introduces issue;
- gives evidence;
- explains mechanism;
- contrasts viewpoints;
- provides example;
- evaluates;
- concludes.
If the question asks for a reason, the explanation paragraph may matter more than the opening example.
11. Question Job Before Answer
| Question job | Answer must do |
|---|---|
| identify | name the relevant item |
| explain | show cause/relationship |
| infer | state implication supported by clues |
| compare | state relevant similarity/difference |
| evaluate | judge using criterion + evidence |
12. Evidence Selection: Use the Smallest Sufficient Set
Students sometimes copy large chunks because they are unsure which detail matters. Train selection:
- What exact claim am I supporting?
- Which phrase proves it?
- Do I need one clue or two?
- What can be omitted without weakening the answer?
13. Paraphrase Without Changing the Meaning
A strong paraphrase preserves the relationship while changing the wording. Students should not replace every word with a synonym; that can distort meaning.
14. Error Families
| Error | Sign | Repair |
|---|---|---|
| literal access | misreads basic event/reference | sentence mapping |
| inference leap | answer plausible but unsupported | evidence-before-inference |
| evaluation drift | personal opinion | criterion→feature→effect |
| multimodal gap | ignores visual evidence | words↔visual comparison |
| answer mismatch | correct idea, wrong question job | verb in question first |
15. A 30-Minute Reading-Depth Practice
| Minutes | Task |
|---|---|
| 0–8 | read for literal meaning/structure |
| 8–15 | reference + evidence questions |
| 15–22 | inference/evaluation |
| 22–27 | multimodal or paraphrase task |
| 27–30 | fresh question-job transfer |
16. How 3-Pax Tuition Can Differentiate Reading Depth
eduKatePunggol’s current format represented on this site is maximum three students, typically 1.5 hours. One passage can expose three different states: one student misses literal reference, another over-infers, while a stronger student needs evaluative precision. The shared text remains efficient while the repair differs.
17. What Not to Do
- Do not infer before literal meaning is stable.
- Do not copy whole paragraphs as evidence.
- Do not treat evaluation as “your opinion”.
- Do not ignore visual information.
- Do not answer an explain question with a label only.
Official 2026 Context
SEAB’s English Language 0001 syllabus states that Paper 2 assesses comprehension of a variety of written and multimodal texts at literal, inferential and evaluative levels. Students should use the current official syllabus and school feedback for their cohort.
Responsible Claims
This is an educational reading framework, not an official answer key or marking rubric. The exact demands of individual Paper 2 items depend on the question and text.
The Main Principle
Go deeper without leaving the text.
Establish the literal base. Connect clues. Infer within bounds. Evaluate against a criterion. Select evidence. Answer the exact job. Deep reading becomes reliable when every step beyond the surface can still show where it came from.
