Primary 4 is where comprehension starts asking children to do more than find words in a passage. A student may read fluently and still struggle because the question now asks for a different kind of proof. The most useful way to teach this change is as an evidence ladder: literal understanding first, then inference, then evaluation.

Quick read

• Literal questions ask what the text directly says.
• Inferential questions ask what the evidence suggests.
• Evaluative questions ask the child to judge, compare or support a position from the text.
• The higher the child moves on the ladder, the more important the evidence-to-explanation link becomes.
• Good tuition should teach children to recognise the demand, find the evidence and explain only as far as the evidence allows.

Why Primary 4 is the right year to make this visible

Lower-primary comprehension often contains many questions where the answer can be located directly. By Primary 4, children increasingly meet less obvious questions. They must connect information across sentences, interpret feelings or motives, explain cause and effect, and support opinions with details from the text.

A child who has learned mainly to search for matching words can therefore appear to “suddenly become weak” in comprehension. The child may still read well. The missing skill is knowing what kind of evidence the question requires.

Level 1: literal evidence

Literal questions ask for information that is directly stated. The main job is accuracy.

• Who did something?
• Where did an event happen?
• What did the character say?
• Which object was used?
• What happened before or after another event?

Even here, students can lose marks by copying too much, missing a condition or selecting the wrong occurrence of a repeated word. The child should learn to identify the answer boundary: enough information to answer fully, but not a whole paragraph of irrelevant text.

Level 2: inferential evidence

Inference asks the child to combine a clue with background understanding. The answer is not copied directly, but it must still be supported.

Suppose a character checks the dark clouds, grabs an umbrella and hurries outside. The text may never say, “She expected rain.” That conclusion is an inference built from evidence.

The useful structure is:

clue from the text → what the clue suggests → answer

Students should not memorise that wording. They should understand the reasoning it represents.

Level 3: evaluative evidence

Evaluation asks the child to make a judgement and support it. A question may ask whether a decision was sensible, which character acted more responsibly, whether a title is suitable or whether a statement is supported by the passage.

The answer now needs both a position and evidence. “Yes, because it was good” is not enough. The child must explain what in the text makes the judgement reasonable.

A useful sequence is:

position → textual evidence → explanation of why the evidence supports the position

The evidence ladder in one example

Imagine a passage where Ben breaks a vase, hides it behind a curtain and later admits what happened after seeing his younger sister being blamed.

• Literal: Where did Ben hide the vase?
• Inferential: Why did Ben hide it?
• Evaluative: Do you think Ben acted responsibly in the end? Support your answer.

The same passage can therefore ask for three different levels of evidence. Children improve when they recognise that difference before writing.

The most common Primary 4 comprehension failures

1. Keyword matching: the child copies the nearest sentence containing a familiar word.
2. Inference without evidence: the answer is plausible but unsupported.
3. Evidence without explanation: the child finds the clue but does not say what it shows.
4. Over-inference: the conclusion goes further than the passage allows.
5. Expression failure: the child understands but writes an incomplete or vague answer.
6. Question failure: the passage is understood but the wrong kind of answer is given.

Teach the question before the answer

Before the child searches the passage, ask: What is this question asking me to do?

• Find?
• Explain?
• Infer?
• Compare?
• Support?
• Judge?

This small pause reduces automatic copying and makes the child more deliberate.

How vocabulary affects the ladder

Vocabulary can break comprehension before reasoning even begins. If the child misunderstands a key verb, emotion word or connector, the evidence may be read incorrectly. Vocabulary therefore needs to be taught in context: meaning, intensity, connotation and sentence use.

The purpose is not to decorate answers with big words. It is to make the text more accurately readable.

How grammar affects the ladder

A child may identify the correct inference but lose clarity through unstable tense, pronoun reference or sentence structure. Grammar therefore has a comprehension role too. The answer must communicate the reasoning accurately enough for the marker to see it.

Changed-passage transfer

After correcting an inference question, do not immediately conclude that the skill has been learned. The same passage still carries memory of the answer. Use a new passage with a different setting and characters but the same reasoning demand.

• New text, same literal boundary problem.
• New character, same motive-inference problem.
• New situation, same evaluative evidence problem.

If the child can recognise and solve the demand again, the learning is travelling.

What a three-student Primary 4 lesson can do

In a three-student, 1.5-hour class, the tutor can give all three children the same passage and compare different reasoning. One child may find the right evidence but over-infer. Another may understand the inference but write vaguely. A third may answer a different question altogether.

The shared text becomes useful because children can hear why one answer is stronger without losing individual correction.

A four-week comprehension cycle

1. Week 1: classify errors by literal, inferential and evaluative demand.
2. Week 2: repair the weakest evidence step.
3. Week 3: use unseen passages and fade prompts.
4. Week 4: compare new independent work with the original errors.

What parents can measure

• The child identifies question types more accurately.
• Copied answers become more selective.
• Inference answers show clearer evidence.
• Evaluative answers contain a position and support.
• Vague “because it is good/bad” explanations reduce.
• The child can answer unfamiliar passages with less prompting.

When tuition may not be necessary

A Primary 4 child who is reading regularly, responding well to school feedback and becoming increasingly accurate across literal, inferential and evaluative questions may not need extra formal support. Parents can continue asking simple evidence questions during reading: “What makes you think that?” and “Which part of the text supports you?”

Related English routes

• Primary 4 English: From Correct Sentences to Coherent Paragraphs
• Primary 4 English Marked-Work Diagnostic
• Primary 5 English Vocabulary Transfer

Frequently asked questions

My child reads fluently. Why is comprehension still weak?

Reading aloud and constructing meaning are different skills. The child may need vocabulary, question decoding, inference, evidence selection or expression support.

Should children copy phrases from the passage?

Sometimes textual evidence is useful, but copying should be selective and purposeful. The child still needs to answer the question and explain the relationship where required.

What is the strongest sign comprehension is improving?

The child can recognise the evidence demand on a new passage and explain an answer without waiting for the tutor to point to the clue.

The main idea

Primary 4 comprehension becomes easier to teach when evidence demands are made visible. Literal questions ask what the text says. Inferential questions ask what the evidence suggests. Evaluative questions ask the child to judge and support. Build the ladder in that order, then test it on unseen passages until the child can climb without prompting.