Text-evidence and inference literacy means noticing what a passage actually says, making reasonable conclusions from its clues, and summarising the central meaning without exaggerating or inventing details. A learner may recognise every word and still answer the wrong question by confusing a literal fact with a plausible guess. Another may write a long ‘summary’ that copies details while losing the writer’s main point.
Ministry of Education V3.0 Reference 186 teaches Singapore English learners, parents and tutors how to separate stated facts, inference, author’s intention, qualifications, selective evidence and concise summaries. It includes a twelve-stage passage-to-claim protocol, a complete original short passage, fully worked literal and inferential questions, a measured summary-compression example, a percentage comparison used only as fictional textual evidence, an error ledger, FAQs and eduKate Depth–Load–Transfer.
This independent eduKateSG article is not an official MOE/SEAB comprehension mark scheme, a prescribed summary word limit or a substitute for school assessment instructions. The fictional passage and figures below were created for this lesson, not collected as actual school data. Age, G1/G2/G3 language level and task requirements matter. Pupils must produce their own assessed answers within the real teacher’s instructions.
1. Reading Comprehension Is a Claim-and-Evidence Skill
Suppose a student reads that someone changed a school library sign after noticing misplaced books. The question asks why the sign was changed. The relevant evidence concerns the initial problem and the person’s decision. Saying the person ‘loves bright colours’ may sound imaginative but contradict the detail that wording mattered more than colour in the particular passage.
Good comprehension begins by finding the actual stated information and identifying the level of inference the text permits. It is not a competition to provide the most complicated psychological theory about a fictional character.
2. Singapore’s Current English Learning Context
MOE’s primary syllabus directory links the Primary English Language syllabus and its reading/viewing learning. SEAB’s G1 2027 directory, G2 directory and G3 directory identify distinct English Language syllabuses. Real reading and summary questions should follow the learner’s particular level, cohort and subject.
NLB’s S.U.R.E. for Schools reinforces the Source–Understand–Research–Evaluate habit, relevant when a passage cites a source or includes a disputed claim. That source-checking framework complements but does not replace the student’s own close reading of the text provided by the school.
3. Eight Rooms in a Reliable Comprehension Answer
| Room | Question | Common hidden failure |
|---|---|---|
| Task | Does the question ask for detail, inference or summary? | Every answer treated as opinion |
| Evidence | Which exact phrase is relevant? | Unrelated quotation used |
| Speaker | Who made this claim? | Character speech treated as narrator confirmation |
| Sequence | What happened before and after? | Earlier plan substituted for final action |
| Vocabulary | Which sense fits the sentence? | Dictionary meaning chosen without context |
| Certainty | Does the author say may, some, must or all? | Qualification erased |
| Compression | Which ideas are essential to the passage? | Minor details dominate the summary |
| Transfer | Can the pupil do the task in a different text? | Model answer memorised |
4. Twelve-Stage Passage, Inference and Summary Protocol
Stage 1 — Identify the actual reading task
Does the question ask what the passage directly states, what can reasonably be inferred, what the writer’s purpose is, or how to summarise the argument? Those demand different evidence. Read the question as carefully as the story itself.
Stage 2 — Locate the relevant part of the passage
Before explaining a character’s motive, find the action, dialogue or descriptive phrase that the inference is based on. An accurate answer about another paragraph will not necessarily address the specific question.
Stage 3 — Separate literal detail from inference
Literal information is explicitly given, while an inference draws a conclusion from clues and relevant context. An inference may be sound without being certain. A reader should say which words make it plausible rather than manufacture a private thought.
Stage 4 — Identify who speaks and when
Direct speech, reported speech and narrator description have different evidential roles. A character’s claim may not be confirmed by the narrator. Track time markers such as ‘earlier’, ‘afterward’ and ‘instead’ to avoid replacing a later decision with an earlier plan.
Stage 5 — Notice contrast and qualification
Words such as but, however, may, some, not always and although can change the scope of an entire statement. A summary that deletes ‘may’ and reports ‘must’ changes the author’s confidence and may become wrong.
Stage 6 — Understand vocabulary in its actual context
A familiar word can have a less familiar sense in a Science, History or literary passage. Use sentence clues, grammatical role and suitable verified sources. The first dictionary entry or AI-generated paraphrase may not fit the author’s meaning.
Stage 7 — Make a tentative inference tied to a clue
For example, a person who double-checks all book titles may value accuracy, but the text may not establish why or whether that trait applies in all settings. Use language proportionate to the available support instead of treating a plausible guess as documented biography.
Stage 8 — Distinguish evidence from a quotation dump
A quotation helps only when the student explains how it supports the claim. A long copied excerpt can hide a missing reasoning link. Follow school citation rules and use the smallest relevant wording needed for the educational task.
Stage 9 — Identify the main idea before writing a summary
Ask what the entire passage is mainly about, not which colourful adjective first attracted attention. Retain essential actors, key change, supporting result and qualification. A summary is selective compression, not a list of every chronological detail.
Stage 10 — Write in genuinely independent words
Preserve the source’s meaning while rephrasing its structure and selecting relevant details. Avoid introducing unmentioned causes, outcomes or emotions. Do not copy an AI-generated summary for an individually assessed task unless it is explicitly permitted and appropriately credited.
Stage 11 — Check every claim against the original
Compare the summary with the passage for numbers, sequence, negation, pronouns, causal claims and scope. If the original says ‘a brief trial was encouraging’, a summary should not say ‘the new system was scientifically proven’.
Stage 12 — Test transfer on a new text
Give another short age-appropriate passage with a different context and ask for one literal answer, one evidence-backed inference and a bounded summary. A child who can do this unaided is showing deeper reading capability than one who memorises answers to a familiar story.
5. An Original Fictional Text for Close Reading
The following is an original, fictional student-library passage written specifically for this teaching reference. It does not describe a real Singapore school, student, data collection or experiment.
On Monday, Nadira placed a cardboard sign beside the library’s return box. The sign said, ‘Please check the book’s title before returning it.’ She had found three books in the wrong section that morning. At lunch, Amir offered to design a brighter sign, but Nadira said the wording mattered more than the colour. They tested the revised sign with ten practice cards: eight were sorted correctly, compared with six for the original. Nadira thanked Amir and wrote that the brief trial was encouraging, though it did not prove the new sign would help every reader.
The teaching passage contains approximately 95 words under a simple word-token count; actual examination word-count conventions may differ. The passage is deliberately small enough for students to verify each claim directly. It includes an early problem, a proposed response, an important contrast and a qualified outcome.
6. Worked Literal Questions: What Does It Say?
| Question | Supported answer | Relevant clue |
|---|---|---|
| What did Nadira place near the return box? | A cardboard sign | First sentence |
| What problem had Nadira noticed? | Three books were in the wrong section | Second and third sentences |
| What did Amir initially offer? | A brighter sign | At lunch |
| What did Nadira value more? | The wording rather than the colour | Her statement |
| How many test cards were sorted correctly with the revised sign? | Eight of ten | Reported fictional practice result |
A literal answer should not be embellished. If the question asks the number of misplaced books, answer three rather than describing a large problem in a busy library. Adding imaginary contextual details may make a story sound richer without making the answer more accurate.
7. Worked Inference Questions: Reason From a Clue
| Possible inference | Text support | How certain? |
|---|---|---|
| Nadira cared about clarity of instructions | She said wording mattered more than colour | Reasonable within this situation |
| Amir wanted to contribute to a solution | He offered to design a sign | Plausible |
| Nadira trusted the revised sign to work everywhere | She explicitly said the trial did not prove that | Unsupported and contradicted |
| The original sign was unattractive | Not stated | Unknown |
| They learned something from a small test | They compared six of ten with eight of ten | Supported as a small fictional comparison |
An inference can be appropriate even if the passage does not contain the answer verbatim. But the reader must distinguish what is plausible from what is certain and demonstrate the bridge between clue and conclusion. A student should not claim to know a character’s entire personality from one decision.
8. The Importance of ‘But’, ‘Though’ and ‘Not’
The phrase ‘but Nadira said the wording mattered more than the colour’ changes the direction of Amir’s proposal. The clause ‘though it did not prove the new sign would help every reader’ limits what the trial established. If the child drops these words, the meaning changes substantially.
| Original textual signal | Effect | Faulty paraphrase |
|---|---|---|
| but | Introduces contrast | Nadira agreed colour was the main issue |
| compared with | Establishes two different recorded counts | Only the revised sign was assessed |
| though | Qualifies the positive result | The trial proved general success |
| did not prove | Explicitly rejects certainty | Success was confirmed for all readers |
| every reader | Defines a universal population claim | Result from ten cards applies to everyone |
9. An Original, Faithful Summary
A defensible short summary is written in the child’s own words and retains the central issue, intervention, comparative result and uncertainty. One possible teacher-created summary is:
Nadira and Amir revised a library-return sign after books were misplaced. In a brief fictional card-sorting exercise, correct placements rose from six out of ten to eight out of ten. Nadira welcomed the result without claiming the sign would help everyone.
The illustrative summary contains approximately 41 words with this token-counting convention. Its length is a demonstration of selective compression, not a mandatory PSLE or SEC word target. The pupil still needs to respond according to the question’s actual allowed length.
10. Why the Summary Is Better Than a List of Details
| Source detail | Include in a short central summary? | Reason |
|---|---|---|
| Nadira noticed misplaced books | Yes | Explains why a change was attempted |
| The sign was cardboard | Usually optional | Material not central to the outcome |
| Amir offered a brighter design | Optional in tight summary | Can be omitted unless design contrast is the focus |
| Correct placements increased from six to eight of ten | Yes where evidence matters | Reports the limited result accurately |
| The trial did not prove universal effectiveness | Yes | Preserves the crucial qualification |
| It happened on Monday at lunch | Often optional | Timing not central unless asked |
A summary should not reproduce every event in equal detail. The aim is to compress the reasoning without deleting the condition that prevents an overclaim. Summarising therefore uses judgement, not merely reducing the number of words.
11. Fictional Percentage Calculation Embedded in the Story
The made-up activity compares six correct placements out of ten practice cards with eight out of ten. Those are 60% and 80% respectively. The difference is 20 percentage points. Relative to the initial 60% rate, the rise is one third (approximately 33.3%); these are different descriptions.
| Fictional sign version | Correct cards | Total cards | Proportion |
|---|---|---|---|
| Original | 6 | 10 | 60% |
| Revised | 8 | 10 | 80% |
| Absolute number difference | 2 | Two more correct placements | |
| Percentage-point difference | 20 percentage points |
These invented results cannot establish that the sign alone caused the change, because order effects, familiarity, group selection and other conditions are unspecified. The passage itself recognises the limitation. English comprehension includes understanding what data support and what they do not prove.
12. A Contradictory ‘Summary’ and Its Repair
Consider this deliberately inaccurate sentence: ‘Amir’s brightly coloured sign made every child return books correctly, proving that colour solved the school’s library problems.’ It changes the subject from wording to colour, invents ‘every child’, ignores the limited trial and turns a small observed difference into proof of causation.
| Fault | What the passage says | Repair |
|---|---|---|
| Says colour solved the issue | Wording mattered more to Nadira | Retain the contrast |
| Says every child | Ten fictional cards were tested | Use the actual narrow sample |
| Says all were correct | Eight of ten were correct | Preserve count |
| Claims proof | Nadira explicitly rejected universal proof | State a positive but limited result |
A useful teacher can ask the child to point to the exact words that would have to exist for the exaggerated sentence to be true. Because they do not appear, it should not be used as a comprehension answer.
13. Inferring Feelings Without Mind Reading
Authors may convey emotion through actions, dialogue, descriptions and context. A character’s behaviour can suggest concern or relief, but one detail rarely establishes a complete private psychological state. The child can use phrasing such as ‘This suggests that…’ when uncertainty is real, while giving a firmer answer when explicit evidence supports one.
Do not mistake qualified wording for weakness. In a text that clearly says ‘She was furious’, reporting anger is a literal reading. In a passage that only describes someone tapping a pencil, anger is only one possibility and may not be the best supported interpretation.
14. Writer’s Purpose and Tone Are Distinct From Literal Event
The fictional passage is written to describe a small attempt to solve an everyday library problem and show the limit of a brief test. It does not announce a national policy or order real students to follow a new procedure. Its tone is matter-of-fact rather than panicked, as shown by measured comparisons and a qualified conclusion.
Reference 182 — Persuasive Language and Tone discusses how writers attempt to influence readers. Reference 186 is narrower: finding valid text evidence, drawing a proportionate inference and compressing meaning accurately. The same skills connect but serve distinct comprehension demands.
15. Source-Based Reading Beyond Fiction
A public article or official notice may contain dates, statistics and quoted statements. Students should identify the original publisher, scope and update date when they need to verify a changing fact. A quoted official statement may describe a policy intention but does not automatically prove its outcomes.
Source evaluation adds another layer to close reading. If the question simply asks what the source claims, answer that accurately. If it asks whether the claim is reliable, examine the independent evidence and relevant provenance. Do not conflate the two tasks.
16. A Reading Error-Diagnosis Ledger
| Incorrect response | Possible gap | Focused repair |
|---|---|---|
| Amir made the cardboard sign at the start | Speaker and sequence confusion | Track who did each action |
| The problem concerned lost books | Word or detail substitution | Reread ‘wrong section’ |
| Nadira preferred brighter colours | Missed ‘but’ | Mark the contrast clause |
| Every reader benefited | Universal overclaim | Preserve ‘did not prove’ |
| The passage is about famous libraries | Main-idea confusion | Name central issue and outcome |
| Summary copies entire passage | Selection problem | Retain only key meaning and limits |
17. A Fifteen-Minute Classroom Text-Evidence Challenge
- Read the passage once for the main event.
- Underline one stated fact and one change-of-plan signal.
- Give a literal answer with the exact relevant detail.
- Write one inference supported by a specified clue.
- Reject an overconfident conclusion and explain why.
- Draft a short summary in new wording.
- Compare the summary against the original for scope and numbers.
- Try a different story with new people, place and outcome.
18. eduKate Education OS: Depth, Load and Transfer
The independent eduKate Education OS separates Depth (understanding claims, inference and qualifiers), Load (answering an appropriate set accurately) and Transfer (using the method on an unfamiliar passage). These are educational diagnostics, not official SEAB English bands.
| Dimension | Student evidence | Possible repair |
|---|---|---|
| Depth | Explains why the trial did not prove a result for every reader | Teach quantifiers and limits |
| Load | Answers varied literal and inferential questions accurately | Use short close-reading practice |
| Transfer | Summarises a different text without inventing claims | Vary genre and evidence |
19. Parent and Tutor Questions
- Which sentence contains the answer?
- Is that explicitly stated or inferred?
- What does the word ‘but’ change?
- Who is speaking and at what point in the story?
- Which detail supports your inference?
- Have you changed ‘may’ into ‘must’?
- What is the passage mainly about?
- Can you remove a minor detail without losing the main claim?
- Does the summary preserve the numbers and qualifications?
- Can the child do the same with a new text unaided?
20. Frequently Asked Questions
Is every reasonable guess an inference?
No. A supported inference follows relevant text clues and context. A guess with no textual basis is not a valid comprehension conclusion.
Must a summary contain every detail from the passage?
No. It should preserve central meaning and essential qualifications while omitting details that do not serve the question.
Can the student copy the passage’s exact wording?
Follow task and citation rules. For independent summary practice, re-express the ideas accurately instead of merely cutting or rearranging the original text.
Does a high percentage in a story prove causation?
Not by itself. A small or poorly controlled comparison may support only a descriptive finding.
Should all inferential answers use ‘maybe’?
No. Calibrate confidence to the evidence. A firmly supported conclusion can use firm language; uncertainty should remain when it genuinely exists.
Are PSLE and 2027 SEC comprehension rules identical?
No. Use the actual assessment and subject-level syllabus, including task-specific limits and marking guidance.
21. Closing Principle
Read what is said. Find the right clue. Infer only what it supports. Preserve qualifications. Summarise without invention. Good comprehension brings evidence and meaning together rather than decorating a guess.
Related eduKateSG References and Primary Sources
- MOE V3.0 Reference Library
- Reference 34 — Source Provenance
- Reference 116 — Disciplinary Reading
- Reference 128 — School Library
- Reference 144 — Summaries and Notes
- Reference 152 — Question Instructions
- Reference 156 — Paraphrasing and Citation
- Reference 160 — Writing and Revision
- Reference 164 — Vocabulary in Context
- Reference 168 — Historical Sources
- Reference 182 — Author’s Purpose and Persuasion
- Immutable Clementi Teaching Reference
- MOE — Primary English Syllabus Directory
- SEAB — 2027 G1 English
- SEAB — 2027 G2 English
- SEAB — 2027 G3 English
- NLB — S.U.R.E. Information Literacy