A photograph may be perfectly genuine while the sentence beside it is too strong. In a school exhibition, a blurred frame, a cropped image and a bold caption raise three separate questions. Can the camera show what the curator says it shows? Does the numerical comparison preserve scale and context? Does the wording allow the reader to draw a defensible conclusion?
This guide brings together three distinct learning jobs: G3 Physics explains how a converging lens forms a real image, Mathematics checks size, proportions and numerical claims, and English examines evidence, audience and effect. It uses imaginary exhibits rather than real student photographs, and it never asks you to decide an individual’s character from an image.
Quick answer: establish the original observation and conditions; identify the transformation; test any calculation against its denominator; rewrite the public claim only as strongly as the evidence warrants. A photograph is a record of a view, not a complete record of an event.
Contents
- How to read a case file
- What the 2027 SEC actually assesses
- The three owners and their boundaries
- A common evidence record
- Physics lens and focus checks
- Mathematics ratio and denominator checks
- English purpose, inference and restraint
- An integrated school exhibition scenario
- Twenty-four worked decisions
- How to repair a claim
- An independent four-week route
- Frequently asked questions
- Sources and next routes
How to read a case file
Begin with what has actually been supplied. The camera operator has an original uncropped frame, an image recorded under particular conditions and a caption prepared for a display. The school may also have measurements, attendance records, event notes or a second photograph. These are different pieces of evidence. Before comparing them, say what each one can establish on its own. This avoids the two opposite errors of trusting every caption and doubting every image.
A practical method is to draw four columns on a page: claim; observation; calculation; qualification. Copy the headline without changing its quantifier. In the observation column state exactly what is visible or measured. In the calculation column show the relevant units, bases and steps, rather than guessing from the size of a printed image. In the qualification column give a sentence that would remain defensible if a skeptical reader asked for the original record.
What the 2027 SEC actually assesses
The Singapore-Cambridge Secondary Education Certificate begins in 2027. Under the official school-candidate G3 syllabuses, English Language is K300, Mathematics K310 and Physics K323. These are the 2027 K-codes, not the old O-Level reference numbers. The SEC does not create a separately marked EMS examination. An EMS investigation is a useful teaching context in which the skills of three subjects remain distinct.
The 2027 K300 English Language assessment has four papers: Writing (35%), Comprehension (35%), Listening (10%) and Oral Communication (20%). A visual text is examined in Paper 2, Section A. K310 Mathematics uses two papers worth 50% each. K323 Physics has multiple-choice, structured/free-response and practical papers weighted 30%, 50% and 20% respectively. The numerical and scientific tasks below are practice scenarios, not claims that an official examiner will show a particular camera question.
The three owners and their boundaries
Physics can explain that a converging lens produces a real, inverted image on a suitably placed screen and that focus changes with object distance. Physics alone cannot prove that a crop honestly represents the whole event. Mathematics can prove that reducing width and height by the same factor preserves shape, or that a proportion derived from selected people says nothing about everyone excluded from the frame. English can explain how the headline, caption and placement direct a reader’s attention. It cannot turn an unsupported statement into a measurement.
Read the supporting route in order: Physics camera focus — Vol 0791; Mathematics camera crop — Vol 0792; English visual-text judgement — Vol 0793. Each article owns its own concept. This page owns the integrated decision; it does not replace the detailed lens explanation, the crop calculations or the visual-text technique.
A common evidence record
Suppose a school displays photographs from a fictional community garden activity. The original landscape frame is 4000 pixels wide and 3000 pixels high. A web card uses a central crop 2000 pixels wide and 1500 pixels high, then exports it at 1000 by 750 pixels. These dimensions are invented for practice. Since both the crop and export preserve a four-to-three shape, their ratios are equal; the first crop still removes three quarters of the original pixel area. Equal proportions do not mean equal coverage.
The curator’s sign says, ‘Everyone in our school planted vegetables.’ The crop shows eight cheerful pupils holding gardening tools. An authentic crop of eight participants might fairly illustrate the activity, but it does not establish whole-school participation. If a verified register separately shows 180 of 240 pupils attended, the attendance rate is 75%, not 100%. The register supports an attendance statement, while the photograph illustrates some participation. These sources should not be silently substituted for one another.
Physics lens and focus checks
An image projected on a screen requires real convergence of light rays. When a converging lens is arranged appropriately, rays from a point on an object meet at the corresponding point on the screen. If the screen or lens is misplaced, the image may be blurred. That is an optical condition, not evidence that the event itself was chaotic. A phone photograph introduces further processes, but the G3 lens model remains a useful foundation for separating the physical image from an editorial claim.
Do not infer that a subject changed size because the projected image became larger. A changed lens-screen arrangement, object distance or display scaling may explain the result. If two photographs are being compared, list camera position, subject distance, orientation, focal setting and crop before making a physical claim. Even a good sharpness test says little about the actual participation rate. A photograph is affected by both the scene and the way it was captured.
Mathematics ratio and denominator checks
The arithmetic is usually simpler than the interpretation. A crop from 4000 by 3000 to 2000 by 1500 has half the width and half the height, so one quarter of the original pixel area remains. An export at 1000 by 750 again halves both linear dimensions and keeps one quarter of the crop’s pixels. The final export contains one sixteenth of the original pixels. The ratio of dimensions is unchanged throughout, yet most original positions are unavailable in the final card.
If the image contains 8 pupils and a list has 12 names, avoid saying two thirds of the school took part. The two thirds would compare visible pupils with a list of 12, assuming the eight are among those twelve. It would not measure the school’s wider population. Every percentage needs a named numerator, a named denominator and a justified comparison group. A proportion without a population is only a decoration.
English purpose, inference and restraint
A visual can appeal to emotion by choosing a close view of smiling participants. An English answer can state this effect without calling the participants dishonest. A caption such as ‘Our garden brings the whole community together’ is persuasive. Ask whether it is clearly an aspiration, a metaphor or a factual claim. Examiners often test a reader’s ability to infer an intended impression and justify it with particular details.
A useful answer pattern is: ‘The image shows [specific visible detail]. This may make the reader feel or think [plausible effect], because [link between detail and purpose]. However, it does not establish [wider unobserved claim].’ The final sentence belongs where evaluation is requested; do not add skepticism to a straightforward literal question. Purpose, audience and wording determine the right amount of analysis.
An integrated school exhibition scenario
A fictional pupil council plans two public posters about the same school garden. Poster A shows an uncropped wide view of 24 pupils at a single session. Poster B shows a tight crop of six pupils and a headline saying, ‘Every class joined the project.’ The publicity team also has a register showing that 9 of 12 classes sent at least one representative. Physics checks whether a blurred area is a focus artifact; Mathematics checks 9 divided by 12; English distinguishes ‘75% of classes were represented’ from ‘every class joined’.
The corrected caption can read: ‘Representatives from nine of the twelve classes joined the first garden session. The photograph shows six of the participants.’ This is less dramatic than the original headline, but it has a durable advantage: the numbers, image and words now agree. A reader can ask about the other three classes without discovering a contradiction. Strong communication is not always the largest possible claim; it is the clearest claim the evidence can carry.
Twenty-four worked decisions
These fictional cases are designed to be answered aloud first and then in writing. The ‘best response’ is not a magic sentence to memorise; it demonstrates where to set the boundary between evidence and a claim. A strong learner names what is established, what remains unverified and what additional information would matter.
1. An authentic close-up
Situation. A display contains a genuine close-up of four volunteers, and the curator says everyone worked happily.
Defensible answer. The close-up establishes that four people were present and may show facial expressions at that moment, not the feelings of every attendee. Why it works: Explain the difference between a local observation and a whole-group emotional claim.
Transfer question. Ask what record, if any, documents the full group and its reactions. When you change the task from an exhibition to a G3 data-response question, keep the evidence and the stated denominator visible. Explain what a new observation would need to show before you strengthen the conclusion.
2. Changing camera distance
Situation. A photograph taken from nearby makes a garden planter look twice as large as in a distant shot.
Defensible answer. Apparent size in the frame depends on geometry and capture conditions; physical enlargement is not established. Why it works: State the variable that changed before trying to compare object dimensions.
Transfer question. Compare only after checking camera placement and a known scale. When you change the task from an exhibition to a G3 data-response question, keep the evidence and the stated denominator visible. Explain what a new observation would need to show before you strengthen the conclusion.
3. Same aspect ratio
Situation. A poster is resized from 1600 by 1200 pixels to 800 by 600.
Defensible answer. Both width and height are halved, and the four-to-three shape is preserved; pixel count becomes a quarter. Why it works: Use multiplicative factors on lengths and square them for area.
Transfer question. Do not describe this export as a crop without evidence that content was removed. When you change the task from an exhibition to a G3 data-response question, keep the evidence and the stated denominator visible. Explain what a new observation would need to show before you strengthen the conclusion.
4. Crop then resize
Situation. A 3000 by 2000 photo is cropped to 1500 by 1000, then exported at 750 by 500.
Defensible answer. The crop retains one quarter of the image’s pixel area and the export one quarter of the crop’s; the final file has one sixteenth the original pixels. Why it works: Separate lost field of view from later changes in digital resolution.
Transfer question. A lower pixel count alone cannot tell which people were outside the crop. When you change the task from an exhibition to a G3 data-response question, keep the evidence and the stated denominator visible. Explain what a new observation would need to show before you strengthen the conclusion.
5. Different denominators
Situation. A photo shows 14 pupils while a sign says 14 of 16 helpers were photographed.
Defensible answer. The supported rate is 87.5% of the listed helpers if all 14 match the list, not 87.5% of all school pupils. Why it works: Identify the reference group explicitly in the answer.
Transfer question. If the question asks school participation, request school-wide data. When you change the task from an exhibition to a G3 data-response question, keep the evidence and the stated denominator visible. Explain what a new observation would need to show before you strengthen the conclusion.
6. Blurred rear row
Situation. Two faces in the back row are blurred and cannot be identified.
Defensible answer. Blur can reflect depth of field, motion or focus, but identity or attendance cannot be established from the blur alone. Why it works: Describe the visual limitation without diagnosing the students.
Transfer question. Do not infer unwillingness or poor conduct from image quality. When you change the task from an exhibition to a G3 data-response question, keep the evidence and the stated denominator visible. Explain what a new observation would need to show before you strengthen the conclusion.
7. Multiple frames
Situation. A contact sheet combines images from three separate garden sessions.
Defensible answer. It may document activity on three occasions, but it is not one moment in which all pictured people stood together. Why it works: Compare timestamps and labels before treating the sheet as a group photograph.
Transfer question. Write ‘across three sessions’ rather than ‘at the same time’. When you change the task from an exhibition to a G3 data-response question, keep the evidence and the stated denominator visible. Explain what a new observation would need to show before you strengthen the conclusion.
8. Overlapping views
Situation. Two photographs show the same table from opposite ends; each contains seven pupils.
Defensible answer. Adding seven and seven would double-count those who appear in both frames. Why it works: Use matching details and unique identifiers, not raw visual totals.
Transfer question. A numerical count needs a rule for duplicates. When you change the task from an exhibition to a G3 data-response question, keep the evidence and the stated denominator visible. Explain what a new observation would need to show before you strengthen the conclusion.
9. Promotional headline
Situation. A cropped picture of a tidy corner appears under the headline ‘Our school is spotless’.
Defensible answer. The image can illustrate one well-maintained location, not prove a condition across all rooms and times. Why it works: Name the overreach in the universal quantifier ‘spotless’.
Transfer question. A safer headline is ‘A tidy corner of our school garden’. When you change the task from an exhibition to a G3 data-response question, keep the evidence and the stated denominator visible. Explain what a new observation would need to show before you strengthen the conclusion.
10. Caption versus register
Situation. A caption claims 100% participation, but a register shows 96 of 120 invited pupils attended.
Defensible answer. The percentage is 80%; the documented count contradicts the universal claim. Why it works: Show the calculation and state the source used.
Transfer question. If the register is partial, avoid even the 80% claim until its scope is established. When you change the task from an exhibition to a G3 data-response question, keep the evidence and the stated denominator visible. Explain what a new observation would need to show before you strengthen the conclusion.
11. A real image
Situation. An illuminated object forms an inverted image on a screen behind a converging lens.
Defensible answer. The image is real because light rays meet at the screen; inversion is an optical property, not evidence that the object was physically turned upside down. Why it works: Use a ray sketch to distinguish object from image.
Transfer question. Do not confuse a digital screen showing an image with a physical convergence test. When you change the task from an exhibition to a G3 data-response question, keep the evidence and the stated denominator visible. Explain what a new observation would need to show before you strengthen the conclusion.
12. Virtual image
Situation. A pupil peers through a magnifier and sees an upright enlarged view that cannot be caught on a screen in that configuration.
Defensible answer. A virtual image can appear larger but is not a projected real image at that apparent position. Why it works: Distinguish observing through the lens from measuring an object.
Transfer question. State the chosen lens arrangement before classifying the image. When you change the task from an exhibition to a G3 data-response question, keep the evidence and the stated denominator visible. Explain what a new observation would need to show before you strengthen the conclusion.
13. Wide-angle surprise
Situation. Near objects appear large relative to far objects in a photograph from a close viewpoint.
Defensible answer. Perspective and distance affect relative appearance; the claim ‘the nearer object has grown’ is unjustified. Why it works: Ask whether the physical objects were measured separately.
Transfer question. The frame alone gives no standard ruler unless one is supplied. When you change the task from an exhibition to a G3 data-response question, keep the evidence and the stated denominator visible. Explain what a new observation would need to show before you strengthen the conclusion.
14. Before and after
Situation. A notice pairs two pictures taken from different heights with different crops.
Defensible answer. A visual contrast may be suggestive but does not isolate the effect of the project. Why it works: Match viewpoint, framing and lighting to strengthen a comparison.
Transfer question. In writing, avoid ‘proved’ when the capture conditions differ. When you change the task from an exhibition to a G3 data-response question, keep the evidence and the stated denominator visible. Explain what a new observation would need to show before you strengthen the conclusion.
15. Hidden footnote
Situation. Large text promises ‘all classrooms transformed’, while small print says ‘sample from two classrooms’.
Defensible answer. The footnote sharply limits the population covered by the claim. Why it works: Explain how size and placement create a stronger initial impression.
Transfer question. The reader should treat qualifications as part of the claim, not an optional extra. When you change the task from an exhibition to a G3 data-response question, keep the evidence and the stated denominator visible. Explain what a new observation would need to show before you strengthen the conclusion.
16. Four-to-three framing
Situation. An image is 1200 by 900 pixels and a space requires 600 by 600.
Defensible answer. A square display cannot preserve the entire four-to-three frame without empty space or distortion; a central square crop removes some sides. Why it works: State the geometric choice as well as its editorial consequence.
Transfer question. Check if the displayed crop omitted a relevant object. When you change the task from an exhibition to a G3 data-response question, keep the evidence and the stated denominator visible. Explain what a new observation would need to show before you strengthen the conclusion.
17. Selectivity at an event
Situation. A photographer deliberately focuses on a winning team rather than the whole tournament.
Defensible answer. The photo fairly represents winners if labelled as winners, but not necessarily the ability of every team. Why it works: Interpret selection in relation to the stated purpose.
Transfer question. The claim becomes unfair only when the selected group is presented as a representative census. When you change the task from an exhibition to a G3 data-response question, keep the evidence and the stated denominator visible. Explain what a new observation would need to show before you strengthen the conclusion.
18. Apparent participation
Situation. One pupil appears in five photographs taken on different days.
Defensible answer. Five appearances establish repeated documentation, not five unique participants. Why it works: Count people, not photos, when the denominator is people.
Transfer question. A clearer count requires identifying unique participants. When you change the task from an exhibition to a G3 data-response question, keep the evidence and the stated denominator visible. Explain what a new observation would need to show before you strengthen the conclusion.
19. Uncertain time
Situation. A photo has no date, and an announcement says a facility ‘now works for everyone’.
Defensible answer. Without a timestamp the image cannot verify the present state; the word ‘now’ requires temporal evidence. Why it works: Identify the time reference missing from the evidence.
Transfer question. A dated inspection record could support the current claim more directly. When you change the task from an exhibition to a G3 data-response question, keep the evidence and the stated denominator visible. Explain what a new observation would need to show before you strengthen the conclusion.
20. A value from a chart
Situation. A bar beneath the photograph shows 9 classes, but the caption claims 9 in 10 classes.
Defensible answer. Nine classes is a count; ‘9 in 10’ is a rate requiring a denominator of ten. Why it works: Check the total before converting a number to a fraction.
Transfer question. If the school has 12 classes, the rate is three quarters. When you change the task from an exhibition to a G3 data-response question, keep the evidence and the stated denominator visible. Explain what a new observation would need to show before you strengthen the conclusion.
21. Causal language
Situation. An exhibition says a new camera caused more people to join the club.
Defensible answer. Photograph quality and club membership are distinct outcomes. A before-after picture of a busy room does not establish causation. Why it works: Ask for enrolment records, time period and alternative explanations.
Transfer question. Use ‘documented attendance’ unless a stronger causal study exists. When you change the task from an exhibition to a G3 data-response question, keep the evidence and the stated denominator visible. Explain what a new observation would need to show before you strengthen the conclusion.
22. Emotion and proof
Situation. A photo of smiling students makes a clean-up day feel successful.
Defensible answer. Facial expressions may support a persuasive mood, but they cannot by themselves establish measured environmental improvement. Why it works: Give a language-effect answer with the exact depicted detail.
Transfer question. Keep the physical outcomes for measurements of waste removed or area cleaned. When you change the task from an exhibition to a G3 data-response question, keep the evidence and the stated denominator visible. Explain what a new observation would need to show before you strengthen the conclusion.
23. Numerical uncertainty
Situation. A printed label rounds an attendance estimate of 74.6% to 75%.
Defensible answer. Rounding may be appropriate, but it must not be interpreted as precise proof that exactly three in four people attended. Why it works: Distinguish a rounded report from the exact count.
Transfer question. Where the count is available, prefer the numerator and denominator. When you change the task from an exhibition to a G3 data-response question, keep the evidence and the stated denominator visible. Explain what a new observation would need to show before you strengthen the conclusion.
24. Final public sign
Situation. A draft reads ‘The camera proves our entire school is engaged’ beside a six-person image.
Defensible answer. Rewrite it to describe six pictured participants and, if independently verified, the relevant count of represented classes. Why it works: The strongest defensible message is both specific and inviting.
Transfer question. Ask one student to check the sign using only the sources displayed beside it. When you change the task from an exhibition to a G3 data-response question, keep the evidence and the stated denominator visible. Explain what a new observation would need to show before you strengthen the conclusion.
How to repair a claim
Start with the fewest words necessary to correct the meaning. Replace ‘everyone’ with ‘some participants’, ‘all 12 classes’ with ‘nine of 12 classes’, or ‘proved that’ with ‘was consistent with’ when the data do not isolate a cause. Do not overcorrect into a different unsupported claim such as ‘the programme probably failed’. A correction should reduce overstatement without introducing speculation.
Work in a disciplined sequence: first verify the source and original frame; next verify the numerical process; then rewrite the sentence for the intended reader. Keep the link between the exact detail and the conclusion explicit. If the exhibition includes an ethical question, respect the privacy of any pictured student. The best interpretation usually needs no claim about personal motives or character.
Four-week independent learning route
Week 1: practise Physics with three lens configurations. Sketch object and image positions; classify real and virtual images; explain why a change in focus is not evidence of a changed event. Write a one-sentence conclusion after each sketch and identify one limitation. Keep the observations separate from editorial interpretation.
Week 2: use ten invented pixel-size records and five attendance tables. Calculate dimension ratios, pixel areas and participation fractions. State the denominator before every percentage. Practise spotting cases where a fraction uses the wrong population or a resize is confused with a crop.
Week 3: read ten fictional poster captions, including deliberately strong quantifiers. Identify a visual detail, the likely effect on the reader and the scope of the factual claim. Rewrite each caption accurately without removing its human interest. Check whether every correction can be justified by the provided text or record.
Week 4: run a full simulated exhibition review. Give yourself an original frame description, crop dimensions, two captions and a register. Produce a Physics explanation, Mathematical working and a public-facing revision. Read the revision aloud to someone unfamiliar with the case. If that reader asks what the evidence means, revise the sentence rather than making it longer.
Frequently asked questions
Does every crop mislead? No. Cropping is often a normal design choice. The test is whether relevant context is hidden and whether the resulting claim is broader than the evidence. Does a blurred picture prove poor scientific method? No. Blur may be caused by an optical arrangement or motion. It can limit what the observer sees, but it does not prove what took place outside the frame.
Is the EMS task a separate SEC paper? No. EMS is a study method that integrates knowledge. In the actual 2027 SEC, English, Mathematics and Physics are examined under their respective G3 syllabuses. Do I need to memorise digital-camera technical specifications? No. Learn the assessed light-and-lens principles and use any additional camera details as context supplied in a question, not as an invented syllabus requirement.
Should I write ‘not enough information’ for everything? No. Say exactly what the evidence does establish first. A careful response can be affirmative about what is visible and qualified about larger claims. What if two sources disagree? Check whether they measure the same people, time, unit or event before deciding which is more relevant.
Sources and next routes
Official 2027 school-candidate SEC subject codes and syllabuses: SEAB G3 syllabuses. See the SEAB explanation of SEC for the examination transition and scheduling. The K300 English, K310 Mathematics and K323 Physics specifications determine actual assessed content; the exhibit cases here are original training material.
For a foundation route, open PSLE English, Mathematics and Science learning guide; Examination Craft route; English, Mathematics and Science Voyage. The practical next move is to redo a single case with a different crop, population or caption and see whether your conclusion remains justified.
