Three correct-looking pages can still tell inconsistent stories
The school optics team is ready to publish its visitor guide. One page explains how the light ray enters a glass block. Another contains a table of incidence angles, refraction angles and calculated sine ratios. A third invites visitors to see a ‘perfect demonstration that all glass bends light identically’. Each page looks convincing. Taken together, they do not yet agree. The last sentence overstates what the first two pages establish.
This guide completes a four-volume sequence with separate owners. Vol 1089 teaches glass-block refraction Physics; Vol 1090 owns sine-ratio and graph calculations; Vol 1091 owns English Paper 2 interpretation and summary. The present volume is not another lesson in the same individual skills. It asks the more advanced question: what should a team publish after checking whether its scientific explanation, measurements and public wording are mutually defensible?
The quick decision rule for a public claim
Identify the exact statement proposed for publication. List what would have to be true for it to hold. Find the original observation, its measured values, the model assumptions and the text in which it appears. Check that every representation uses the same boundary, direction of travel, angle reference and level of certainty. If a chain breaks, revise the claim instead of defending it with extra adjectives.
Then issue one of three decisions: publish as supported, publish with qualification, or withhold pending better evidence. This is a classroom audit method for learning, not a formal professional certification. The final statement should be understandable to a visitor who did not watch the experiment and cannot see the raw notebook. Calm clarity is a form of accuracy.
Do not invent an EMS examination code
EMS in this series means a combined English–Mathematics–Science learning exercise. It is not a separately registered Singapore-Cambridge SEC subject. Under the SEAB 2027 G3 directory, English Language is K300, Mathematics K310 and Physics K323. A student taking combined Science may be entered for a different combination and code. Do not transfer K323 paper details to everyone taking Science.
For the individual official syllabuses, see K300 English, K310 Mathematics and K323 Physics. K300 includes comprehension, multimodal reading and summary; K310 tests mathematical modelling and communication; K323 covers refraction, sine ratios and practical investigations. Their assessment structures remain distinct. The SEC starts in 2027 and records subjects at their taken levels; this learning project does not replace any official paper.
Define a claim that can actually be tested
A display might propose: ‘Our measurements are consistent with a refractive-index ratio near 1.50 for light entering the classroom block from air at the recorded incident angles.’ This is testable. The notebook should identify the setup, normal, rays and measurements needed to evaluate it. By contrast, ‘We have proved light behaves perfectly in all transparent objects’ changes the question. The student team has not tested all media, wavelengths or geometries.
A useful audit starts with the nouns and qualifiers. Which block? Which direction of travel? Which light source? Which measured angles? Which model relation? Which uncertainty? Each answer narrows the statement until the evidence can carry its weight. Do not mistake a narrow, accurate statement for an uninteresting one. A well-designed display can be engaging without promising more than it knows.
Build a small evidence register before writing the poster
The team lists five items: the block outline, a ray-tracing record, the original angle table, a calculation sheet and a draft poster caption. Each entry should have a clear label or trial identifier. The outline explains the reference surface. The ray record shows direction. The table preserves raw readings. The calculation sheet explains how derived values were produced. The draft caption shows what the audience is being asked to believe.
Provenance means knowing where a record came from and how it was handled. An ideal diagram from a textbook may help explain the model but cannot be presented as the observed trace of today’s block. A neat spreadsheet cannot replace the rough original data if the cells were later edited. A revised poster needs a record of the claim it replaced, especially when a qualification has been added. This discipline can be practised with paper folders rather than special software.
Audit the normal before judging an angle
The normal is drawn perpendicular to a surface at the point of incidence. If the surface changes orientation, so must the normal. The audit therefore begins with the photograph or outline, not the sine table. An incidence angle measured from the block face instead of from the normal can produce a plausible-looking number that is physically mislabelled. Mathematics cannot correct an incorrectly defined variable by being precise.
The reviewer should reconstruct one trial independently. Mark the incident ray, the point where it meets the boundary, the local normal and the transmitted ray. Ask a second student to repeat the construction without seeing the first student’s angle labels. Only then compare values. A few minutes spent on geometry can save an entire page of invalid calculations.
Follow direction of travel through both faces
In the simple air-to-glass part of the model, the ray bends towards the normal at oblique incidence. When it leaves a parallel-sided block into air, it bends away. An ideal emergent ray is parallel to the incident ray when the outer media match and the two faces are parallel, but it may be shifted sideways. An exhibit cannot claim it is precisely the same line merely because the directions are parallel.
A reviewer should check arrowheads and name both boundaries. An image without arrows invites the wrong interpretation. A second face that is not parallel changes the geometry and may invalidate the parallel-emergence claim. If the exhibit uses a triangular object, do not quietly borrow the rectangular-block conclusion. This check is basic Physics, but in an evidence review its purpose is to verify that all published representations describe the same apparatus.
The data table is not the conclusion
Consider a fictional set of classroom records: incident 30 degrees, refracted 19.5 degrees; incident 40 degrees, refracted 25.4 degrees; incident 50 degrees, refracted 30.7 degrees. The values are illustrative, not actual school measurements. They are roughly consistent with a sine ratio near 1.50 if the relevant angles are measured to the normal and the approximate air-to-glass model applies.
The reviewer must check those conditions before announcing the ratio. A single mislabelled 30-degree angle could reverse the interpretation. The values shown to one decimal place also should not be treated as infinitely precise. The calculation sheet may reasonably estimate a trend, but the public statement should not claim that each reading proves an identical exact constant. It should say what the synthetic example demonstrates and, in a real project, what actual readings show.
A graph’s gradient depends on its axis labels
If a graph plots sin i vertically and sin r horizontally, a best-fit slope around 1.50 is consistent with the selected ideal model. Reverse the axes and the slope becomes approximately 0.667. The numbers differ because gradient always means change in the vertical variable divided by change in the horizontal variable. The exhibit should not label either result ‘refractive index’ without explaining the interpretation.
An audit compares the graph’s axes, calculated column headings and accompanying caption. If the caption claims 1.50 but the plot displays the reciprocal relation, the team may still have correct data and correct algebra; its explanation is what needs repair. The solution is to name the axes, state the relationship and give the interpreted quantity. This is precisely the kind of cross-document error a good reviewer detects.
Check the rounded figures against the words
A calculated index shown as 1.50 suggests three significant figures in the mathematical display, but it should not automatically lead to the words ‘exactly 1.50000’. In a real practical, the instrument’s limitations matter. A wide light beam, thick pencil markings, a shifted outline and imperfect normal placement can all affect angle readings. A responsible caption reports an estimate rather than an absolute material property unless the evidence warrants more.
The K310 Mathematics syllabus generally expects non-exact numerical answers to three significant figures and angle answers to one decimal place unless specified otherwise. Those are answer-format conventions. They do not make experimental evidence precise beyond its measurements. The review distinguishes mathematical reporting style from uncertainty about the physical observation. Treating those two matters as identical is a common source of misleading school display language.
Keep contradictory trials visible
Suppose another fictional trial gives incidence 55 degrees and refraction 36 degrees. Its sine ratio is about 1.39, noticeably lower than the other illustrative values near 1.50. Do not delete the row merely because it spoils the message. Ask whether the refracted angle was taken from the normal, whether the block shifted, whether the same source and boundary were used, and whether the result can be repeated.
An outlier is a reason to investigate, not a licence to invent an explanation. The team may report that one reading was inconsistent with the cluster and that further checks are needed. If a documented transcription error is discovered, record the correction transparently. A reader should not have to choose between believing a polished graph and believing the raw notebook. The two should be reconciled honestly.
An English claim can fail while the Physics remains sound
The poster says ‘Every transparent object bends every beam in the same way’. This is too strong even if the glass-block calculation is correct. The Physics concerns a particular interface and optical conditions; the sentence expands to every object and beam. The appropriate editing move is to narrow the language, not to accuse the experimental team of bad science. A small wording change can repair a major inference error.
The English reviewer circles universal quantifiers such as ‘all’, ‘every’, ‘never’, ‘always’ and ‘perfect’. For each, ask which evidence would have to exist to support that scope. Where none does, replace the word with a condition or identify the particular test. This is a disciplined language habit also useful for comprehension, persuasive writing and science-news reading.
A qualification must be close enough to be useful
A notice may present the headline ‘Glass bends light’ and hide ‘at an oblique interface under these conditions’ in a distant footnote. Technically the footnote might supply the missing qualification; practically a visitor may miss it. The audit should consider visual hierarchy, not only whether correct words occur somewhere on the page. A correction that nobody sees has limited communication value.
A clearer caption reads: ‘In this demonstration, a ray entering our glass block at an oblique angle changed direction towards the normal.’ A second sentence can state that the angle pairs were consistent with an approximate sine relationship. The reader can now distinguish direct observation from quantitative interpretation. The phrase ‘in this demonstration’ is not timid; it is a precise statement of what was actually shown.
Avoid confusing a model drawing with a video record
A textbook-style ray diagram typically assumes an ideal smooth interface and a single thin ray. A classroom video may show a wider band of light, a bright reflected component and a slightly irregular trace. These two representations are useful for different purposes. One clarifies the expected rule; the other documents what was seen. The exhibition must label them so that nobody mistakes an ideal drawing for raw evidence.
If a screenshot has been cropped, indicate that it is an extract when that matters to interpretation. If the drawing uses a calculated ray path, say it is a model. Do not imply an animation was captured by the experiment merely because it looks realistic. The task is not to ban illustrations but to preserve the distinction between explanation, simulation and observation.
A defensible panel needs three different voices
The Science student asks whether the physical account is valid: which media, which boundaries, which ray direction, and whether the stated cause is correct. The Mathematics student checks angles, sines, axes, calculations and rounding. The English student asks whether the intended audience can understand the text and whether every claim reflects the evidence. Each reviewer signs off on a different kind of correctness.
The group should also challenge one another. A mathematically correct gradient may be attached to a physically wrong reference angle. A scientifically accurate caveat may be expressed so vaguely that readers miss it. A beautiful concise sentence may omit the condition that makes it true. Collaborative review is valuable only when the participants compare the same claim and agree on what is and is not established.
Use a claims ledger rather than a collection of opinions
The simplest ledger has columns for claim, evidence, assumption, concern and decision. For ‘the refraction angle was smaller than the incident angle in these trials’, the notebook and diagram may provide support. For ‘the ratio was exactly the same in all observations’, the rounded table and scatter may not. For ‘the ray came out parallel to the incident ray’, the rectangular block geometry matters. The required evidence changes with the claim.
Each decision should cite a record, not a student’s confidence. A reviewer who writes ‘looks convincing’ has not completed the audit. Another who writes ‘not enough evidence’ should specify what is missing: a normal, a repeat, a scale, a direction arrow, a graph axis or a clear qualification. Precision about the missing piece turns criticism into useful correction.
How to respond when two reviewers disagree
Suppose the Physics reviewer accepts the diagram and the Mathematics reviewer rejects the angle table. Do not take a simple vote. Identify the exact point of disagreement. Perhaps the ray paths are conceptually correct but the angles were measured from the wrong face. In that case the diagram can remain as an explanatory model, while the numerical claim requires remeasurement. One part of the work need not invalidate every other part.
Similarly, an English reviewer may request a clearer headline without disputing the calculated result. Resolve disputes at the smallest affected unit. This protects worthwhile work and teaches students to avoid all-or-nothing judgements. A strong audit produces a concrete correction path rather than a vague complaint about quality.
Publish, qualify or withhold
Publication decisions can be simple. Publish a claim when its evidence, assumptions and wording align. Publish with a qualification when the central finding is supported but the scope or precision needs narrowing. Withhold when the relevant evidence is missing or contradictory in a way that prevents the conclusion. A school exhibition can openly state that a question remains unresolved; that is better than inventing an answer.
The final visitor guide might say: ‘Our classroom ray traces were broadly consistent with the expected direction change for oblique air-to-glass refraction. The numerical estimate is approximate because the angles were read from drawn lines.’ It should not say: ‘This proves every ray in every glass follows exactly our measured value.’ The difference is not marketing style. It is the boundary between observation and overclaim.
Recover the PSLE-to-SEC habit of evidence before conclusion
Primary English teaches accurate retrieval and attention to what the text says. Primary Mathematics teaches labelled quantities, reasoned working and sense checks. Primary Science teaches observation, explanation and fair comparison. The SEC version of the same discipline requires better control over sources, inference and models. A learner does not need to start from nothing; earlier habits can be upgraded.
For foundations, see the local guides to PSLE English, PSLE Mathematics word problems and PSLE Science explanations. For the secondary subject bridge, Sengkang’s G3 preparation route provides context. These are links to distinct owners, not duplicate versions of the current evidence-audit guide.
An advanced learner checks the counterexample
One diagram may show a ray incident along the normal, with no visible direction change. The broad poster claim that glass always visibly bends light fails on this straightforward counterexample. A careful student asks what feature of the original conclusion remains true: light may change speed on crossing an interface even though its propagation direction does not change at normal incidence. The counterexample repairs an over-general statement without denying refraction as a physical process.
Counterexamples are powerful in all three subjects. In Science they reveal missing conditions. In Mathematics they test whether a proposed rule applies universally. In English they expose the danger of unrestricted words such as ‘always’. This is a higher-level examination habit: find the smallest case that would make a claim fail, then rewrite the claim with its proper condition.
A good review is transparent but not exhausting
The final panel need not publish every line of the laboratory notebook. It should state the method at an appropriate level, present a few representative readings with clear labels and make the underlying records available to a teacher or organiser where relevant. Visitors need a coherent route through the evidence, not a wall of numbers. Transparency means that the claim can be checked, not that every raw detail must appear in the headline.
Separate the public explanation from the technical appendix. The public caption says what happened and under which conditions. A small data note identifies the angle basis and approximate relationship. A teaching note can explain the calculation. Each layer serves a different reader. Good modern nonfiction honours that distinction rather than treating length as a substitute for clarity.
Eighteen cross-subject evidence hearings
Hearing 1: The model presented as data
A slide prints a perfectly smooth graph generated from the equation, then labels it ‘our measured results’.
Panel decision: The label is inaccurate. Call the graph an ideal model, or replace it with a plot of actual readings and retain any scatter.
Challenge question: Which record could establish whether the graph came from a real trial? Explain whether the correction belongs to the physical model, the mathematical representation, the language or more than one of these. Naming the owner prevents a vague ‘redo everything’ response.
A useful written conclusion has three steps: identify the particular evidence, specify the claim it licenses, and give an action proportional to the weakness. If only a caption is overstated, edit the caption. If the raw angle reference is missing, recheck the observations. Do not treat all defects as interchangeable.
Hearing 2: A hidden normal
The ray trace gives 30 and 20 degrees but no perpendicular reference line.
Panel decision: Withhold the numerical interpretation until the normal and angle definitions are recovered. The numbers may refer to the surface instead.
Challenge question: What simple diagram correction would make the result audit-ready? Explain whether the correction belongs to the physical model, the mathematical representation, the language or more than one of these. Naming the owner prevents a vague ‘redo everything’ response.
A useful written conclusion has three steps: identify the particular evidence, specify the claim it licenses, and give an action proportional to the weakness. If only a caption is overstated, edit the caption. If the raw angle reference is missing, recheck the observations. Do not treat all defects as interchangeable.
Hearing 3: A swapped ray arrow
The caption says the ray enters from air, but the arrows show it travelling from glass to air.
Panel decision: The Physics statements and drawing disagree. Correct the direction or rewrite the media and bending explanation after checking the actual setup.
Challenge question: Which claim cannot be accepted until the direction is settled? Explain whether the correction belongs to the physical model, the mathematical representation, the language or more than one of these. Naming the owner prevents a vague ‘redo everything’ response.
A useful written conclusion has three steps: identify the particular evidence, specify the claim it licenses, and give an action proportional to the weakness. If only a caption is overstated, edit the caption. If the raw angle reference is missing, recheck the observations. Do not treat all defects as interchangeable.
Hearing 4: The accurate reciprocal
A plotted slope is 0.667 because y is sin r and x is sin i, yet a text says ‘our ratio is 1.50’.
Panel decision: Both can describe the same model when the reciprocal relationship is explained. Show the axes and state the transformation clearly.
Challenge question: What would make the numerical caption misleading even if both numbers are correct? Explain whether the correction belongs to the physical model, the mathematical representation, the language or more than one of these. Naming the owner prevents a vague ‘redo everything’ response.
A useful written conclusion has three steps: identify the particular evidence, specify the claim it licenses, and give an action proportional to the weakness. If only a caption is overstated, edit the caption. If the raw angle reference is missing, recheck the observations. Do not treat all defects as interchangeable.
Hearing 5: The stray 1.39
One fictional result gives a sine ratio near 1.39 while three others cluster around 1.50.
Panel decision: Investigate the outlier; do not erase it for presentation. The audit may keep the cautious central finding with a note about variability.
Challenge question: Suggest one realistic check of tracing or angle recording. Explain whether the correction belongs to the physical model, the mathematical representation, the language or more than one of these. Naming the owner prevents a vague ‘redo everything’ response.
A useful written conclusion has three steps: identify the particular evidence, specify the claim it licenses, and give an action proportional to the weakness. If only a caption is overstated, edit the caption. If the raw angle reference is missing, recheck the observations. Do not treat all defects as interchangeable.
Hearing 6: A claim about all glass
The exhibit says the experiment measured every type of glass.
Panel decision: That is unsupported by one school block. Narrow the claim to the sample actually observed and the conditions recorded.
Challenge question: Which universal word should be removed first? Explain whether the correction belongs to the physical model, the mathematical representation, the language or more than one of these. Naming the owner prevents a vague ‘redo everything’ response.
A useful written conclusion has three steps: identify the particular evidence, specify the claim it licenses, and give an action proportional to the weakness. If only a caption is overstated, edit the caption. If the raw angle reference is missing, recheck the observations. Do not treat all defects as interchangeable.
Hearing 7: Parallel versus same line
The model shows the emergent ray parallel to the incident ray but offset sideways.
Panel decision: Say parallel, not identical or coincident. A lateral displacement does not contradict equal directions in the ideal parallel-faced model.
Challenge question: What extra geometry is needed to discuss displacement magnitude? Explain whether the correction belongs to the physical model, the mathematical representation, the language or more than one of these. Naming the owner prevents a vague ‘redo everything’ response.
A useful written conclusion has three steps: identify the particular evidence, specify the claim it licenses, and give an action proportional to the weakness. If only a caption is overstated, edit the caption. If the raw angle reference is missing, recheck the observations. Do not treat all defects as interchangeable.
Hearing 8: An angle to the face
A learner finds 70 degrees on the drawing, measured to the interface.
Panel decision: Convert to 20 degrees to the normal for the standard sine relationship. Keep the raw note, and show the corrected variable definition.
Challenge question: Why could an apparently precise calculation still fail? Explain whether the correction belongs to the physical model, the mathematical representation, the language or more than one of these. Naming the owner prevents a vague ‘redo everything’ response.
A useful written conclusion has three steps: identify the particular evidence, specify the claim it licenses, and give an action proportional to the weakness. If only a caption is overstated, edit the caption. If the raw angle reference is missing, recheck the observations. Do not treat all defects as interchangeable.
Hearing 9: A dramatic headline
The poster uses ‘Perfect Light-Bending Test’ above a cautious data note.
Panel decision: The language implies more certainty than the note supports. Replace the headline with a compelling but conditional statement.
Challenge question: Draft a six- to ten-word headline that does not overclaim. Explain whether the correction belongs to the physical model, the mathematical representation, the language or more than one of these. Naming the owner prevents a vague ‘redo everything’ response.
A useful written conclusion has three steps: identify the particular evidence, specify the claim it licenses, and give an action proportional to the weakness. If only a caption is overstated, edit the caption. If the raw angle reference is missing, recheck the observations. Do not treat all defects as interchangeable.
Hearing 10: An unclear unit
A number is called 1.50 degrees in the exhibit.
Panel decision: The refractive-index-related ratio is dimensionless in the simple calculation. Correct the label without changing the angle data.
Challenge question: Which quantities in the supporting table should carry degree units? Explain whether the correction belongs to the physical model, the mathematical representation, the language or more than one of these. Naming the owner prevents a vague ‘redo everything’ response.
A useful written conclusion has three steps: identify the particular evidence, specify the claim it licenses, and give an action proportional to the weakness. If only a caption is overstated, edit the caption. If the raw angle reference is missing, recheck the observations. Do not treat all defects as interchangeable.
Hearing 11: A copied figure
A diagram was copied from a class handout but its caption says ‘photograph of our trial’.
Panel decision: Revise the caption to distinguish illustrative model from observation. A handout figure cannot establish this experiment’s result.
Challenge question: What would an authentic observation record include? Explain whether the correction belongs to the physical model, the mathematical representation, the language or more than one of these. Naming the owner prevents a vague ‘redo everything’ response.
A useful written conclusion has three steps: identify the particular evidence, specify the claim it licenses, and give an action proportional to the weakness. If only a caption is overstated, edit the caption. If the raw angle reference is missing, recheck the observations. Do not treat all defects as interchangeable.
Hearing 12: The normal-incidence test
A beam enters perpendicular to the face and passes straight through in the ideal diagram.
Panel decision: This is consistent with refraction at normal incidence and contradicts the headline ‘light always changes direction’. Revise the claim.
Challenge question: Which physical quantity can change even when direction does not? Explain whether the correction belongs to the physical model, the mathematical representation, the language or more than one of these. Naming the owner prevents a vague ‘redo everything’ response.
A useful written conclusion has three steps: identify the particular evidence, specify the claim it licenses, and give an action proportional to the weakness. If only a caption is overstated, edit the caption. If the raw angle reference is missing, recheck the observations. Do not treat all defects as interchangeable.
Hearing 13: Technical jargon
A caption contains six specialist nouns and no clear statement about the observed beam.
Panel decision: Replace jargon with a short observation, then define the normal and the conditional model in a second sentence. Accuracy and accessibility can coexist.
Challenge question: Which technical term is essential rather than ornamental? Explain whether the correction belongs to the physical model, the mathematical representation, the language or more than one of these. Naming the owner prevents a vague ‘redo everything’ response.
A useful written conclusion has three steps: identify the particular evidence, specify the claim it licenses, and give an action proportional to the weakness. If only a caption is overstated, edit the caption. If the raw angle reference is missing, recheck the observations. Do not treat all defects as interchangeable.
Hearing 14: The lonely measurement
The whole exhibit cites one angle pair and calls its result an exact universal law.
Panel decision: One pair may illustrate a calculation but does not show repeatability across conditions. Qualify the finding and suggest more trials.
Challenge question: How would an expanded table improve the strength of the claim? Explain whether the correction belongs to the physical model, the mathematical representation, the language or more than one of these. Naming the owner prevents a vague ‘redo everything’ response.
A useful written conclusion has three steps: identify the particular evidence, specify the claim it licenses, and give an action proportional to the weakness. If only a caption is overstated, edit the caption. If the raw angle reference is missing, recheck the observations. Do not treat all defects as interchangeable.
Hearing 15: A lab-review disagreement
The Science member approves the causal diagram; the English member rejects the public wording.
Panel decision: Keep the valid explanation, revise the overbroad words and document the change. The reviewers are evaluating different dimensions.
Challenge question: What issue would justify rerunning the physical experiment rather than merely editing? Explain whether the correction belongs to the physical model, the mathematical representation, the language or more than one of these. Naming the owner prevents a vague ‘redo everything’ response.
A useful written conclusion has three steps: identify the particular evidence, specify the claim it licenses, and give an action proportional to the weakness. If only a caption is overstated, edit the caption. If the raw angle reference is missing, recheck the observations. Do not treat all defects as interchangeable.
Hearing 16: A minor decimal
A public caption reports 1.498732 as the property of the block, although angles were drawn approximately.
Panel decision: Use a sensible rounded estimate and acknowledge practical limits. Detailed calculator output is not a measurement instrument.
Challenge question: Which figure should be retained for internal calculations but not public certainty? Explain whether the correction belongs to the physical model, the mathematical representation, the language or more than one of these. Naming the owner prevents a vague ‘redo everything’ response.
A useful written conclusion has three steps: identify the particular evidence, specify the claim it licenses, and give an action proportional to the weakness. If only a caption is overstated, edit the caption. If the raw angle reference is missing, recheck the observations. Do not treat all defects as interchangeable.
Hearing 17: A plausible but unsupported inference
Someone claims all visitors will be confused because one student misread the sign.
Panel decision: The student report suggests a readability issue but does not establish the experience of every visitor. Recommend an additional pilot check.
Challenge question: Write the finding using ‘may’ or ‘some’ in a defensible way. Explain whether the correction belongs to the physical model, the mathematical representation, the language or more than one of these. Naming the owner prevents a vague ‘redo everything’ response.
A useful written conclusion has three steps: identify the particular evidence, specify the claim it licenses, and give an action proportional to the weakness. If only a caption is overstated, edit the caption. If the raw angle reference is missing, recheck the observations. Do not treat all defects as interchangeable.
Hearing 18: The publish decision
The team has a labelled model, a traceable angle table, a corrected graph and a qualified caption.
Panel decision: Publish the limited claim with its conditions and an invitation to examine the method. State remaining uncertainty without implying the exhibit is useless.
Challenge question: List the minimum evidence needed for final sign-off. Explain whether the correction belongs to the physical model, the mathematical representation, the language or more than one of these. Naming the owner prevents a vague ‘redo everything’ response.
A useful written conclusion has three steps: identify the particular evidence, specify the claim it licenses, and give an action proportional to the weakness. If only a caption is overstated, edit the caption. If the raw angle reference is missing, recheck the observations. Do not treat all defects as interchangeable.
A complete audit walkthrough with invented records
Begin with the proposed caption: ‘Our glass block makes every ray bend by the same amount.’ Underline ‘every’ and ‘same amount’. Next examine the ray diagram. The oblique rays change direction differently at different incident angles, and a normal-incidence example supplies a counterexample to visible bending at every angle. The Physics reviewer therefore rejects the unconditional sentence.
Now inspect the mathematical record. The imaginary angles 30/19.5, 40/25.4 and 50/30.7 degrees give similar sine ratios near 1.50, but not equal differences in angle measured in degrees. The Mathematics reviewer explains that a near-constant ratio of sines is not a constant amount of bending in degrees. This distinction is the precise hidden error in the original caption. A mathematically informed rewrite should not replace one false universal sentence with another.
The English reviewer proposes: ‘For several oblique rays entering our rectangular glass block from air, the observed directions were broadly consistent with the expected refraction pattern. Our angle measurements suggest an approximate sine ratio near 1.50.’ The team checks that ‘observed’ is reserved for real trials; if the displayed table is only generated model data, the wording must instead say ‘our model calculations’. That final distinction prevents the invented exercise from being presented as an actual result.
Independent practice for the following week
First, reconstruct the direction arrows and normals from an unlabelled rectangular-block sketch. Second, check one sine ratio without using a formula card. Third, reverse the graph axes and explain why the slope changes. Fourth, edit five public claims by reducing unsupported certainty. Fifth, prepare a short evidence register with source labels. Sixth, hold a group review where each student speaks for one subject. Seventh, complete the audit alone using a new context, such as a different transparent block or a school poster on sound.
Record errors by their origin: physical assumption, measured variable, calculation, graphical representation, source inference or public wording. Revising a whole 5,000-word article will not fix a learner who always measures from the surface; a single diagnostic exercise can. Concentrated feedback followed by a new-context retest is the route from recognition to independence.
Frequently asked questions and final route
Is EMS an actual examination subject? No. It is the English, Mathematics and Science integration used in this educational series.
Must the team publish all raw data on the poster? No. It should preserve evidence and make the public statement traceable without overwhelming the reader.
Can a calculation be correct while the conclusion is wrong? Yes. The selected variables or the wording can contradict the physical interpretation.
Does a single anomalous reading disprove the model? Not automatically. Investigate the method and the pattern; neither conceal the anomaly nor overreact without analysis.
Why is this volume distinct from Vol 1088? That volume audited ripple-tank wavefront reflection. This one audits optical refraction through glass, numerical sine ratios and the claims of a separate exhibit. The systems are connected by careful evidence reasoning, but their physical models differ.
What should I read next? Students who struggle with the physical explanation should revisit Vol 1089. Those who misread graph slopes should practise Vol 1090. Those who over-infer a writer’s claims should revisit Vol 1091. Use the official SEC overview for examination administration and each current syllabus for its subject’s assessment requirements. This independent teaching guide is not an official mark scheme or paper.
