Checked: 2 September 2026. This reader guide is aligned to MOE’s English Language Syllabus 2020 Primary. This legacy “The Flow” URL does not claim a current eduKateSG branch or a school relationship.
English tasks increasingly ask children to read more than a paragraph. A prompt may combine instructions, bullet points, pictures, labels, tables, headings, notices or layout. Students can understand every individual piece and still fail because they never combine those pieces into one task model. For families around The Flow and elsewhere in Singapore, this old 2017 URL now owns one precise English job: how to read text, image and layout together before answering.
Quick read
- Start with the instruction: what output is required?
- Inventory every source of information—words, pictures, labels, headings and layout.
- Separate what is directly shown from what is inferred.
- Identify relationships across the different parts of the prompt.
- Only then select information and construct the response.
- The goal is one integrated task model, not several disconnected observations.
Why multimodal prompts are difficult
A learner may read the written paragraph well but ignore the diagram. Another notices the picture but misses one label. A third copies every visible detail without deciding which ones matter.
The difficulty is integration.
The task may distribute information across:
- instructions;
- body text;
- captions;
- images;
- speech bubbles;
- tables;
- timelines;
- headings;
- graphic layout.
The reader has to combine them while still respecting the task.
Step 1: read the instruction before studying the stimulus deeply
Ask:
What am I supposed to produce?
Possible outputs include:
- a written message;
- a comprehension answer;
- a comparison;
- a summary;
- an oral response;
- a paragraph;
- a decision with reasons.
The instruction tells the learner which information will matter.
Step 2: inventory the information channels
Do a quick scan:
- What text is present?
- What images are present?
- Are there labels?
- Does position or sequence matter?
- Are there dates, times, arrows or categories?
- Is any information repeated?
This prevents one visually prominent element from dominating the whole task.
Step 3: separate observation from inference
If a picture shows a child holding an umbrella under dark clouds, the direct observations are:
- the child has an umbrella;
- the sky appears dark;
- the setting is outdoors.
“It is about to rain” may be a reasonable inference, but it is still an inference unless other information confirms it.
Students should know which layer they are using.
Step 4: read labels as part of the meaning
A photograph without labels and the same photograph labelled “Before” or “After” can support different interpretations.
Likewise, a table heading may determine whether a number represents:
- time;
- cost;
- number of participants;
- percentage;
- ranking.
Labels are not decoration. They constrain meaning.
Step 5: identify relationships across modes
Ask:
- Does the image illustrate the text?
- Does it add information not stated in the text?
- Does the table contradict or qualify a claim?
- Does the layout show sequence?
- Does a caption change how the image should be interpreted?
The task model is built from these relationships.
Step 6: decide what information the response actually needs
Not every visible item belongs in the answer.
A good selection rule is:
If this detail disappeared, would my response become incomplete or less accurate for the task?
If not, it may be low-value.
Pictures in writing prompts
When images support writing, students should avoid two extremes:
- ignoring the image entirely;
- describing every visible object.
Instead ask:
- What idea does this picture make available?
- What conflict, setting, action or relationship could it support?
- Which details are useful to the chosen story?
Visual information becomes raw material, not a checklist.
Pictures in comprehension
When a question refers to both image and text, students should identify which claim comes from which source.
Example:
Text: says the event begins at 4 p.m.
Poster: shows registration closes at 3.45 p.m.
The reader may need both facts to give a complete response.
Tables and schedules
Students often read entries but miss the table’s structure.
Before answering:
- read row labels;
- read column labels;
- check units;
- identify comparison points;
- note blanks or exceptions.
A number without its row and column meaning is incomplete information.
Headings create hierarchy
A heading tells the reader what the following information belongs to.
In notices, webpages or brochures, hierarchy may be shown through:
- size;
- bold text;
- spacing;
- boxes;
- bullet groups.
Students should learn to read layout as organisation rather than treating every line equally.
Sequence can be visual
A timeline, numbered set of images or arrows may show sequence more clearly than prose.
Ask:
- What happened first?
- What changed?
- Which step depends on the previous one?
This is useful in informational and procedural texts.
The whole-prompt paraphrase
Before answering a complex prompt, the learner should be able to say one sentence such as:
I need to write to my teacher to explain which activity I chose, when it happens and why I need permission, using the notice and timetable.
If the learner cannot paraphrase the whole task, they may still be reading fragments.
Multimodal reading and Situational Writing
Situational Writing often requires students to transform stimulus information into a natural message.
The learner should:
- identify audience and purpose;
- extract all required content points;
- combine text/image/layout information;
- write for the recipient;
- check whether the reader could act correctly.
The stimulus is a source, not a sentence template.
Multimodal reading and oral communication
A visual stimulus can support oral discussion, but students should distinguish:
- what they see;
- what it suggests;
- what personal or general knowledge they add.
This keeps elaboration grounded while allowing natural conversation.
Multimodal reading and school notices
Real school life includes:
- timetables;
- forms;
- notices;
- posters;
- maps;
- online pages.
Reading these successfully is part of functional English. The learner needs to extract the action, deadline, location, conditions and audience quickly.
Common failure: image first, instruction second
A striking image captures attention and the student invents an answer before reading the task.
Repair:
Instruction first. Stimulus second. Response third.
Common failure: every detail treated as equally important
Repair:
Rank details by task relevance.
Highlight only information that changes the answer.
Common failure: inference presented as fact
Repair:
Use language such as suggests or appears when the evidence is indirect, and avoid adding unsupported details.
Common failure: layout ignored
Repair:
Read headings, boxes, sequence and labels explicitly before interpreting individual sentences.
Common failure: source copied into awkward prose
Repair:
Extract information into notes, close the source where practical, then generate natural language appropriate to the audience.
Three students and one multimodal task
In a 3-pax class, students can first build individual task models from the same prompt.
Then compare:
- which information each selected;
- what one student missed;
- which inference was unsupported;
- how the layout changed interpretation.
Discussion happens after individual reading so the tutor can see each learner’s original intake.
Catch Up, Keep Up, Move Ahead
Catch Up: use simple posters or notices with explicit labels.
Keep Up: integrate images, text and tables in current school tasks.
Move Ahead: compare how different modes reinforce, qualify or contradict one another and evaluate source choices.
The learner should become increasingly independent.
A 90-minute multimodal lesson
A useful session can:
- read one instruction;
- inventory stimulus modes;
- separate observation/inference;
- build one task-model sentence;
- select relevant content;
- write or answer independently;
- check the response against every required source.
The final question is not “Did I use the picture?” but “Did I integrate the information needed for the task?”
How parents can help
- Ask the child what the instruction requires before discussing the image.
- Ask which information comes from text and which from visuals.
- Ask whether an idea is observed or inferred.
- Use real notices, menus, maps and schedules as occasional reading practice.
- Avoid supplying an interpretation before the child builds their own task model.
Current official source
MOE’s English Language Syllabus 2020 Primary treats listening, viewing, reading, speaking, representing and writing as connected language activities and emphasises effective communication across contexts.
What we removed from the old 2017 page
The historical page referred to “The Flow Primary School” and a Marina Bay tuition centre without a verified basis, used stale contact information and repeated generic service copy.
This rebuild does not preserve those claims. It assigns the URL a truthful educational job: integrating text, image and layout into one task model.
Boundary and locality
This guide may be useful to families around The Flow and elsewhere in Singapore. It does not claim a current eduKateSG branch at The Flow or Marina Bay and does not claim an affiliation with any school. Current class locations are on the contact page.
Frequently asked questions
Should students always describe pictures in a writing task?
No. Use only visual details that serve the task or chosen story.
How can a child tell observation from inference?
Observation is directly visible or stated; inference is a conclusion drawn from those clues.
Why do students miss information in notices?
They may read lines but not headings, labels or layout relationships. Teach the whole-prompt scan.
Does multimodal reading matter beyond exams?
Yes. Everyday English includes forms, schedules, maps, websites, posters and messages that combine several information modes.
What is the long-term goal?
A reader who can turn a visually complex prompt into one coherent mental model and then produce exactly the response that model requires.
The larger point
A prompt is not a pile of parts. Strong readers see the instruction, words, images and layout converge into one problem.