The answer first: compare worst cases, not only the printed measurements
A rigid reference disc is reported as 12.45 mm in diameter and a slot as 12.50 mm wide, both correct to the nearest 0.01 mm. On those assumptions alone, the disc is smaller even in the worst allowed rounding case: its diameter is below 12.455 mm, while the slot is at least 12.495 mm wide. The nominal gap is 0.05 mm and the conservative model gap is at least 0.040 mm. A correct fit decision compares the largest allowed object with the smallest allowed opening, not merely 12.45 against 12.50.
This G3 SEC Mathematics guide explains how to turn rounded measurements into defensible bounds, use intervals in geometry and arithmetic, and state exactly what a model does and does not guarantee. The examples use fictional rigid discs, calipers, holes and exhibition parts. They are original learning problems, not official examination questions or engineering certification. The central habit is simple: a number printed to two decimal places may represent a range, and the decision has to work across the appropriate range.
The linked Volume 1157 Physics practical guide taught the physical work of taking and correcting caliper and micrometer readings. Volume 1154 showed how dimensions govern a regular circle grid, while Volume 1156 examined how calculation becomes a public claim. This volume owns a different problem: how uncertainty in a reported dimension changes the arithmetic and whether a design can be said to fit with justified confidence.
The quick guide to measurement bounds
- Identify the rounding step and take half of it to form the basic rounding interval.
- For 12.45 mm to the nearest 0.01 mm, the ideal interval is [12.445, 12.455) mm.
- Apply any stated exact additive correction to both interval endpoints as well as the centre.
- For a guaranteed physical fit in a model, compare the object’s upper size with the opening’s lower size.
- For positive dimensions, preserve their order when converting units, multiplying or using increasing area formulas.
- Do not treat display rounding, instrument accuracy and manufacturing tolerance as the same source of uncertainty.
- Report whether your result is unconditional, model-dependent or still awaiting information.
Official G3 SEC Mathematics scope
The official 2027 SEC G3 syllabus directory lists Mathematics as K310, with 4052 shown as the earlier reference code. The 2027 K310 syllabus describes two papers, each lasting two hours 15 minutes, carrying 90 marks and contributing 50% of the assessment. The last question of Paper 2 focuses on the application of Mathematics in a real-world situation. Numerical accuracy, units, interpretation, mathematical reasoning and coherent working are therefore important.
The extended interval comparisons in this article are an enrichment route that develops those modelling and precision habits, not a claim that a specific caliper-and-slot proof is explicitly prescribed as an examination topic. An official examination question may give particular accuracy instructions or define an interval differently; follow those instructions. Unless a task specifies another convention, the syllabus gives guidance on appropriate rounding, including three significant figures for many non-exact values and one decimal place for angles. A calculated bound should not be rounded in a direction that silently destroys the guarantee it was meant to provide.
A measurement is often an interval disguised as a number
Suppose a caliper display is read to the nearest 0.01 mm and the reported value is 12.45 mm. That value can represent an unrounded measurement from 12.445 mm up to but not including 12.455 mm under the conventional nearest-hundredth rounding model. The interval’s width is 0.010 mm, and its half-width is 0.005 mm. The upper endpoint notation makes clear that numbers at 12.455 mm would ordinarily round to 12.46 mm under the chosen convention.
Students sometimes take the final displayed decimal place as an error of ±0.01 mm. That doubles the rounding range in this example. The half-step rule is about rounding, not about all physical uncertainty. It is also a mistake to claim that the instrument is accurate within ±0.005 mm merely because it shows two decimal places. Zero offsets, calibration, pressure, alignment and specimen variation may be much larger. Always identify whether the task defines a rounded value, a measured accuracy, or a manufacturing tolerance.
Use words for the direction of an inequality
Lower bound means a value no less than the specified lower endpoint. Upper bound means a value no greater than an upper limit under the model, with endpoint inclusion governed by the rounding rule. If a reported mass is 3.05 g to the nearest 0.01 g, the ideal rounding interval is [3.045,3.055) g. If a manufactured part has a stated tolerance of 3.05 ± 0.02 g, its allowable specification interval is [3.03,3.07] g. The two intervals answer different questions and should not be blended.
First apply the zero correction in the proper order
Recall the Physics measurement case: a digital caliper shows +0.03 mm when closed and 12.48 mm across a disc. Treat the zero offset as an exactly known additive constant for this purely mathematical example. Corrected display centre = 12.48 − 0.03 = 12.45 mm. The pre-correction rounding interval [12.475,12.485) becomes [12.445,12.455) mm when 0.03 is subtracted from both endpoints.
A more realistic measurement may include uncertainty in the zero check. Then it would be wrong to pretend the 0.03 mm offset is infinitely exact. One can use a larger interval if the problem supplies one, or state that a full physical guarantee is not established. The interval model should fit the information given. Do not invent an unstated instrument calibration certificate to make the arithmetic sound more secure.
Compare the limiting pair for a guaranteed fit
A disc fits through a rigid slot only if its actual relevant outside size is no greater than the slot’s actual clear opening. For a proof that works throughout the stated intervals, the worst case is a large disc and a small slot. With the disc’s ideal bounds [12.445,12.455) mm and the slot’s [12.495,12.505) mm, the limiting comparison is 12.455 against 12.495. There is a positive difference of 0.040 mm at those limiting endpoints. The design is guaranteed to fit under that rounding-only, rigid-geometry model.
If the two reported central dimensions were both 12.45 mm, however, the intervals would overlap. The disc could be bigger than the slot. Identical labels do not prove actual equality, and they certainly do not prove positive clearance. If the opening is 12.44 mm to the nearest 0.01 mm, its upper bound is 12.445 mm, while the disc’s lower bound begins at 12.445 mm. Apart from boundary conventions, there is no positive fit clearance supported by these values. The recommended decision changes when the intervals change.
Measure the correct extremum of an irregular object
Some discs are not perfectly round. The Physics practical guide asks students to measure several orientations because the largest external diameter may control whether the object passes through a round opening. Averaging a narrow dimension with a wider one can produce a mathematically valid mean that is physically irrelevant to the fitting task. An answer can be numerically correct yet solve the wrong question. State first whether fit requires a maximum, an average or a particular named axis.
Combine intervals for more than one object
Consider two discs placed side by side, each reported as 12.45 mm in diameter to the nearest 0.01 mm, with a minimum required clear gap of exactly 0.20 mm. The largest total model width is less than 2×12.455 + 0.20 = 25.110 mm. If the opening reads 25.12 mm to the nearest 0.01 mm, its smallest allowed value is 25.115 mm. There is still a positive worst-case margin of at least 0.005 mm under these assumptions. That narrow margin may be unsuitable in a real design if other tolerances are unknown, but the stated rounding model supports the inequality.
Change the opening’s reported width to 25.10 mm and its lower bound becomes 25.095 mm. Now there are permitted combinations whose total disc-and-gap width exceeds the opening. A guarantee is impossible on the data alone. A weaker conclusion—“The nominal numbers look close”—is not a substitute for checking the worst case. The same logic applies to a line of boxes, tiles, cupboards, printed cards or components in a school project.
Use monotonic formulas for area and volume bounds
If a positive diameter d ranges from 12.445 mm to just under 12.455 mm, radius r=d/2 ranges from 6.2225 mm to just under 6.2275 mm. A circle’s area is A=πr². For positive r, area grows as radius grows, so the lower area is π×6.2225² mm² and the upper is π×6.2275² mm². There is no need to guess which way the inequality moves. It follows from the fact that all radii in the problem are non-negative.
For a right circular cylinder with positive radius r and positive thickness h, volume is V=πr²h. To obtain an upper bound using independent positive intervals, use the upper radius and upper thickness. For a lower bound, use their lower endpoints. If one variable is not independent of the others, the relationship must be respected rather than combining impossible extremes. This is a more demanding form of modelling than merely entering a formula in a calculator.
Conversion of units also applies to bounds
Dividing the diameter interval [12.445,12.455) mm by ten gives [1.2445,1.2455) cm. When converting a volume, use the cube of the unit factor: 1 cm³ equals 1000 mm³. A cylinder volume of about 390 mm³ corresponds to 0.390 cm³. Both the central value and its range must be converted on the same basis. A converted central value with unconverted endpoints is internally inconsistent.
Percentage differences require a specified base
Suppose a nominal diameter is 12.50 mm and the corrected central estimate is 12.45 mm. The difference is 0.05 mm. As a percentage of the nominal diameter, the difference is (0.05 ÷ 12.50)×100 = 0.400%. As a percentage of the corrected diameter, it would be (0.05 ÷ 12.45)×100, a different value. These are not contradictory results; they answer different definitions of “percentage difference”. In examinations, name the base before calculating.
Bounds can also affect percentages. If the disc diameter and the nominal reference each have allowed ranges, the minimum and maximum percentage changes depend on how those intervals compare. An answer based on centres alone is a point estimate rather than a guaranteed range. State which one the question requests. Precise mathematical language prevents a fair calculation from becoming a misleading claim.
Twenty worked decision drills
Each drill changes one assumption or mathematical operation. The purpose is to learn a procedure that survives a new context rather than memorise a single disc size. Try to write the relevant interval or inequality before reading the answer. All dimensions are invented training data and the guarantee language is limited to each case’s stated model.
Decision drill 1: Find the bounds of a rounded diameter
Given information: A rigid disc diameter is reported as 12.45 mm correct to the nearest 0.01 mm.
Worked reasoning: In the ideal rounding model, the underlying measured value is at least 12.445 mm and less than 12.455 mm. The half-step is 0.005 mm. A safe interval is [12.445, 12.455) mm. The upper endpoint is excluded under a conventional rounding rule, although when modelling physical tolerances a non-strict upper bound can also be used conservatively.
What to learn: First identify the reported unit and rounding step; do not apply ±0.01 mm when the last displayed place is hundredths of a millimetre.
Decision drill 2: Correct a positive zero offset before making an interval
Given information: A caliper reads +0.03 mm when closed and 12.48 mm on the object, both displayed to 0.01 mm. Treat the stated offset as an exact additive correction in this exercise.
Worked reasoning: Corrected centre = 12.48 − 0.03 = 12.45 mm. The interval associated with rounding the 12.48 mm reading is [12.475,12.485); after subtracting the exact 0.03, it becomes [12.445,12.455) mm. A physical zero check could have its own uncertainty, so this model is deliberately simplified.
What to learn: Apply a correction to both interval endpoints, not merely the displayed centre, whenever the question asks for a range.
Decision drill 3: Correct a negative zero offset
Given information: A different instrument shows −0.02 mm at zero and 12.43 mm on the disc; use the same exact-offset model.
Worked reasoning: Corrected centre = 12.43 − (−0.02) = 12.45 mm. The signed formula avoids a direction error. Its rounded input interval [12.425,12.435) becomes [12.445,12.455) after adding 0.02.
What to learn: The sign is part of the quantity. The instrument reference is not a reason to change the algebraic rule halfway through a problem.
Decision drill 4: Compare a disc with a wider slot
Given information: The disc diameter is reported as 12.45 mm and the slot opening as 12.50 mm, each correct to the nearest 0.01 mm and subject only to that simplified rounding uncertainty.
Worked reasoning: The disc’s upper bound is 12.455 mm and the slot’s lower bound is 12.495 mm. Even the smallest allowed slot exceeds the largest allowed disc by at least 0.040 mm in this rounding-only model. Therefore fitting is supported by these particular intervals, without implying real fabrication accuracy.
What to learn: For a guaranteed-fit claim compare the smallest allowed opening with the largest allowed object, not the two midpoints.
Decision drill 5: Recognise that equal nominal widths do not guarantee passage
Given information: The slot and the disc both read 12.45 mm to the nearest 0.01 mm.
Worked reasoning: Their ideal rounding intervals overlap completely. The slot could be slightly smaller than the disc, or slightly larger. Equal centres do not guarantee clearance. The correct conclusion is that further information or a different specification is required.
What to learn: When a question asks ‘must fit?’, answering from a nominal equality is a common trap. Test the entire allowed intervals.
Decision drill 6: Two discs and a specified gap
Given information: Two discs are each reported as 12.45 mm to 0.01 mm and must have an exact model gap of at least 0.20 mm. An opening reads 25.12 mm to 0.01 mm.
Worked reasoning: The greatest total width demanded by the discs and gap is less than 2×12.455 + 0.20 = 25.110 mm. The smallest allowed opening is 25.115 mm. Under the stated rounding-only assumptions, the two discs fit with a worst-case clearance above 0.005 mm. Real tolerances may change the decision.
What to learn: For several objects, add the upper bounds of the sizes and the required minimum gaps, then compare with the lower bound of available space.
Decision drill 7: The same arrangement with a narrower opening
Given information: Keep the two 12.45 mm discs and 0.20 mm gap, but the opening is now reported as 25.10 mm.
Worked reasoning: The opening may be as narrow as 25.095 mm. The greatest model width needed is just under 25.110 mm. Since some permitted combinations would not fit, the available evidence cannot guarantee assembly. An optimistic nominal comparison would miss that failure.
What to learn: A design can appear to fit at rounded centres while failing a worst-case requirement.
Decision drill 8: Different orientations change the decision
Given information: Measurements of a rigid disc at three orientations read 12.45, 12.49 and 12.51 mm.
Worked reasoning: An average of about 12.48 mm may describe the sample, but fit through a circular opening is controlled by the maximum relevant outside dimension, not necessarily the mean. The evidence also suggests the disc may be non-round. A clearance decision needs the appropriate geometric maximum and measurement uncertainties.
What to learn: Before applying formulas, ask whether a mean, maximum, minimum or representative value actually matches the physical decision.
Decision drill 9: From diameter bounds to radius bounds
Given information: The diameter is in [12.445,12.455) mm and radius is half the diameter.
Worked reasoning: Divide both positive endpoints by two. Radius lies in [6.2225,6.2275) mm. It is not reasonable to give radius the same ±0.005 mm rounding interval as the diameter, because the division scales the interval width.
What to learn: A positive linear transformation preserves the interval order; a scale factor also scales the uncertainty range.
Decision drill 10: Propagate a positive interval into circular area
Given information: Assume the disc is circular and its radius is between 6.2225 and 6.2275 mm.
Worked reasoning: Since area A=πr² increases with positive radius, the lower model area is π×6.2225² and the upper is π×6.2275², in mm². Do not square a rounded central area and assume its error is still the original length error. The unit becomes square millimetres.
What to learn: Identify whether a formula increases or decreases with the varying quantity, then choose the corresponding endpoints for safe bounds.
Decision drill 11: Convert a small volume properly
Given information: A simple cylindrical sample has an estimated volume of 390 mm³.
Worked reasoning: One cubic centimetre equals 1000 cubic millimetres. Therefore 390 mm³ = 0.390 cm³. A division by ten would be the length conversion, not a cubic-volume conversion. This matters when combining a calculated geometry volume with mass in grams to obtain density in g/cm³.
What to learn: Write the power on every unit and convert powers accordingly; the original length unit is not the final volume unit.
Decision drill 12: Separate rounding interval from manufacturing tolerance
Given information: A product specification permits 12.45 ± 0.02 mm, while a measurement is printed as 12.45 mm to the nearest 0.01 mm.
Worked reasoning: The stated specification interval is [12.43,12.47] mm. The ideal rounding interval around the printed measurement is [12.445,12.455) mm. These are not the same information. The specification describes allowed manufactured variation; the second describes one display’s rounding, not all errors in the physical measurement.
What to learn: Never replace an explicit tolerance with half a display step. Use the source of uncertainty actually stated by the question.
Decision drill 13: Work out a percentage shortfall
Given information: The measured central diameter is 12.45 mm; the nominal design diameter is 12.50 mm. Find the central estimate of percentage shortfall relative to nominal.
Worked reasoning: Difference = 12.50 − 12.45 = 0.05 mm. Divide by the stated comparison base 12.50 mm and multiply by 100: 0.400%. State that this is a central estimate. If asked for a guaranteed bound on percent difference, use appropriate endpoint comparisons instead.
What to learn: The percentage denominator is a choice given by the question. ‘Percentage smaller than nominal’ and ‘percentage larger than measured’ have different bases.
Decision drill 14: Use an interval after unit conversion
Given information: The diameter lies in [12.445,12.455) mm; the drawing needs centimetres.
Worked reasoning: Divide each endpoint by ten to obtain [1.2445,1.2455) cm. The interval follows the same conversion as its centre. A sketch labelled 12.45 cm would be wrong by a factor of ten.
What to learn: Apply unit changes consistently to central values, lower bounds and upper bounds. Do not convert only the headline number.
Decision drill 15: Compare a density estimate with appropriate caution
Given information: A cylindrical disc model has mass 3.05 g and calculated volume 0.390 cm³.
Worked reasoning: The central density estimate is 3.05 ÷ 0.390 ≈ 7.82 g/cm³. But if the mass and volume have intervals, the lowest possible density in a positive-data model comes from the smallest mass divided by the largest volume; the highest comes from the largest mass divided by the smallest volume.
What to learn: For a quotient of positive quantities, numerator and denominator changes influence the result in opposite directions.
Decision drill 16: A narrow range does not remove zero bias
Given information: Three raw caliper readings are 12.48, 12.50 and 12.49 mm, and a stable +0.03 mm offset is known.
Worked reasoning: The raw mean is 12.49 mm. Corrected readings are 12.45, 12.47 and 12.46 mm, so corrected mean is 12.46 mm. Both sets have a 0.02 mm range, because subtracting a common constant shifts the entire set without changing its spread.
What to learn: A small spread measures internal consistency. It does not establish absence of systematic offset.
Decision drill 17: State a model limit in the answer
Given information: The caliper resolution is 0.01 mm, and the task asks whether the fit is absolutely guaranteed in real life.
Worked reasoning: A rounding-only interval can prove a statement within the model. It cannot by itself guarantee real-world fit if device calibration, object deformation, shape or actual manufacturing tolerances remain unknown. An accurate conclusion is ‘guaranteed under the stated rounding and geometry assumptions’, followed by a brief qualification if appropriate.
What to learn: The strength of a conclusion should match the stated information. A proof inside a model is not automatically a universal physical guarantee.
Decision drill 18: Integers and clearances do not mix through casual rounding
Given information: Five markers fit a width condition, but a formula with changed measurements returns a value of 5.3 possible positions.
Worked reasoning: The integer capacity is at most five complete positions, assuming the derived inequality is correct. Rounding 5.3 to the nearest whole number could produce five here, but the correct reason is that a sixth object violates a constraint. In other cases ordinary nearest-integer rounding would give an impossible answer.
What to learn: For counts under an upper-capacity restriction, test the largest feasible integer rather than treating a decimal as a measurement to round normally.
Decision drill 19: Two uncertain values and a claim of equality
Given information: The school report calls two rounded diameters ‘identical’ because each shows 12.45 mm.
Worked reasoning: The display values are identical, but their unrounded values need not be. One could be near 12.445 mm and the other near 12.455 mm within their allowed intervals. More properly, they are indistinguishable at the reported rounding step; that is not a proof of perfect equality.
What to learn: In Mathematics and Science, distinguish equality of representations from equality of the underlying physical quantities.
Decision drill 20: Make a final design recommendation
Given information: A committee must buy a slot fitting a rigid disc, and the only measurements are rounded sizes with partial calibration information.
Worked reasoning: Write the dimensional inequality for the actual decision, identify whether rounding and manufacturing tolerances have been supplied, and choose a clearance policy that works for the worst-case allowed values. Where calibration uncertainty or object non-roundness is unknown, request a better specification or measurement rather than making an absolute promise.
What to learn: A real-world modelling answer is strong when it gives a defensible decision, a numerical reason and the key limitation without exaggeration.
A mathematical solution should make its decision auditable
When showing a fit decision, use a predictable sequence. Identify the required fit condition in words: the largest possible outside dimension must not exceed the smallest permitted inside opening. Write the interval for each rounded or toleranced quantity. If a zero correction is given, apply it before comparing. Compute the limiting gap. End with a conditional conclusion: “The disc fits under the stated rounding-only model,” or “The supplied data do not guarantee fit.” Those few sentences show a chain that another reader can inspect.
Some students use the word ‘approximately’ as permission to skip the decision. It is useful for describing the centre or a rounded derived quantity, but it is not enough when the question asks whether a condition is definitely satisfied. Where a threshold is at stake, calculate with appropriate extra internal precision and decide the inequality before rounding the reported value. Rounding a value to a pleasing number is not a reason to reverse a pass/fail verdict.
A four-stage plan for stronger G3 problem solving
Stage one: control the meaning of a rounded number
Start with ordinary units: 7.3 cm to the nearest 0.1 cm, 7.30 cm to the nearest 0.01 cm and 7 cm to the nearest centimetre. These represent progressively different ranges. Ask learners to place each on a number line and identify the half-step. They should recognise that 7.3 and 7.30 can be numerically equal while communicating different rounding information. This foundation prevents errors before more complex measurement contexts appear.
Stage two: use simple additions and subtractions
Combine two uncertain lengths, and subtract one from another. When two intervals are added, the lower total uses lower endpoints and the upper total uses upper endpoints. In subtraction, the smallest possible difference uses the first quantity’s lower endpoint minus the second quantity’s upper endpoint. Students should explain this with a physical example, such as remaining space on a shelf, before relying on rules alone.
Stage three: include geometric decisions
Move to a hole and a disc, then two objects with a gap, then a regular row inside a tray. Ask for one example where comparing central readings is sufficient for a rough estimate and one where a guaranteed-fit decision needs bounds. A student who can explain why those tasks are different is moving from calculation into modelling. Use Volume 1154 for the earlier fixed-dimension grid; do not assume its exact count remains valid when dimensions become uncertain.
Stage four: transfer to unfamiliar examination contexts
Use a new data set describing a school stage backdrop, a container, a sports-equipment case or a packing plan. Change the units and the tolerance statement. Ask learners to decide whether the information proves feasibility, proves impossibility or leaves the result unresolved. Require a short sentence explaining the limiting comparison. If the learner can solve without a hint about which bound to use, the method has become transferable.
Common examination errors and practical repairs
A frequent mistake is to convert 12.45 mm into 12.45 cm or to divide a cubic unit by ten. The repair is to write the conversion factor with units and apply its power. Another is to treat the entire display increment as the rounding half-width; the repair is to mark the two neighbouring midpoints on a number line. A third is to use the mean diameter where the largest possible diameter controls fit; the repair is to state the physical decision first. The most subtle error is to call a rounding-only proof an unconditional real-world guarantee. The repair is to name the assumptions alongside the conclusion.
Frequently asked questions
Are measurement intervals just another rounding trick?
No. Rounding is an entry point, but interval reasoning helps students interpret limits and constraints. It distinguishes a central estimate from a range of possibilities. The same habit is useful when a word problem contains a price cap, safety margin, maximum allowed distance, minimum qualifying score or required capacity. The relevant question is always which extremum controls the decision.
Is every caliper question officially examinable under Mathematics K310?
No specific caliper-and-slot scenario is being promised. The linked Physical Science practical provides a natural cross-subject context, and this Mathematics guide develops modelling and precision habits. Refer to the official K310 syllabus for the actual assessed content. A useful private practice problem is not proof that the national examination will contain that topic.
What if the numerical intervals overlap?
Overlapping ranges often mean the evidence does not establish a guaranteed ordering. It may still be possible for the physical items to fit in some combinations. That is different from proving fit in every permitted combination. State which outcome the question requests: possible, guaranteed, impossible or unresolved. Those words have distinct mathematical meanings.
How do students moving from PSLE reach this level?
Secure measurement units, fractions, decimals, percentages, ratios and problem representation first. The PSLE Mathematics error-diagnosis owner and the PSLE-to-G3 Sengkang transition guide are starting points. A student does not need advanced vocabulary to state that the largest disc must be smaller than the smallest slot; the vocabulary becomes easier after the physical idea is clear.
Final independent challenge
A different caliper reads +0.02 mm at zero and 18.64 mm across a rigid disc to the nearest 0.01 mm. A slot is reported as 18.65 mm to the nearest 0.01 mm. Treat the zero offset as exactly known for this exercise. The corrected disc centre is 18.62 mm, with ideal rounding interval [18.615,18.625) mm. The slot lies in [18.645,18.655) mm. Comparing the greatest disc with the smallest slot leaves at least 0.020 mm clearance under the rounding model. The decision is supported by a chain of reasoning, not by comparing uncorrected 18.64 and 18.65.
For an advanced extension, change the assumption: suppose the zero reading has an unknown calibration error, or the disc is not perfectly circular. Then the proof is incomplete until those new possible variations are bounded. A correct mathematician is willing to change a conclusion when a changed premise requires it. That is a stronger skill than merely carrying more decimal places.
Related owners and verification
Continue with Volume 1157 Physics Measurement Bench, Volume 1154 Mathematics Circular Grid, Volume 1156 EMS Evidence Review, Secondary 1 Data and Representation, and the official 2027 SEAB G3 syllabus directory. All examples, clearances, object labels and measurements here are original instructional data. The national examination’s questions, exact apparatus and future paper content must not be inferred from them.
