A voltage graph is not an argument until the numbers have been checked
Three students test a magnet passing through a coil. A data logger records short voltage pulses. One student sees a negative value and calls it a failure; another takes one large positive value and declares the device twice as powerful; a third draws a straight line through a tiny dataset and predicts that the same relationship will continue at any speed. All three have noticed something real, but none has yet performed a complete mathematical analysis. This Learner’s Guide teaches you to make a disciplined decision from signed values, repeats, ratios and graph models.
Mathematics begins with the meaning of the data. A signed sensor reading uses a reference direction. The number +20 mV and the number -20 mV have equal absolute magnitude, though opposite polarity. A mean describes the centre of repeat measurements, not the exact result of every future run. A linear model describes a possible trend over a measured domain, not an unconditional law of nature. These distinctions connect school algebra and statistics to the scientific practices developed in Vol 0500, the magnet-and-coil induction practical.
Our purpose is narrower than the general electricity or generator unit. Vol 0331 already owns alternating-current generator waveform interpretation. Vol 0497 owns motor rotation counts and rates. This new lesson concentrates on a measured short pulse: which values can be compared, what fair ratios mean, how we distinguish a measured result from a line drawn through it, and how to protect the conclusion from unjustified certainty.
The five-step mathematical method
First identify the quantity, unit, sign convention and measurement context. Second inspect the raw trials rather than trusting a tidy published mean. Third compute the required descriptive values: mean, median, range, percentage difference or ratio. Fourth decide what graph or algebraic model is justified by the data and its domain. Fifth translate the mathematics back into a cautious sentence that says what was measured and what remains unknown. In a problem-solving examination, the final interpretation can be as important as the arithmetic.
Never make the first step a calculator operation. If a table contains peak induced voltage in millivolts, a number without mV has lost its physical meaning. If a table contains direction-coded voltages, an average of signed values may answer one question but not another. If the equipment recorded pulse magnitude instead of polarity, its measurements might all be positive. A correct method has to fit the quantity actually stated in the question, rather than borrowing a familiar formula before reading the headings.
This guide uses fictional apparatus readings to practise mathematical reasoning. They are deliberately neat enough for hand calculation and irregular enough to test judgement. No set of constructed numbers is evidence about a commercial generator, a safety rating, or a universal physical relationship. Students should be able to tell a reader which numbers were supplied, which values were calculated from them and which conclusions would still require additional experiments.
The official SEC examination context
The Singapore–Cambridge Secondary Education Certificate begins in 2027. The official SEAB 2027 G3 school-candidate directory lists Mathematics as K310, with 4052 as the 2026-and-earlier reference code. The K310 syllabus assesses standard techniques, problem-solving in context, and reasoning or mathematical communication. Paper 1 and Paper 2 each last 2 hours 15 minutes, are worth 90 marks and carry 50% of the subject total. Paper 2 ends with a question focused on applying Mathematics in a real-world situation.
The official syllabus says that omitting essential working can lose marks and that, unless otherwise specified, non-exact numerical answers should be given to three significant figures, with angles in degrees usually to one decimal place. The learner must still read the particular question for an overriding instruction. This article is a teaching exercise, not an official paper or guaranteed question type. Its science setting is drawn from the 2027 K323 G3 Physics induction topic, not from a separate EMS examination.
English is K300 under the same 2027 directory. Students need clear language to explain a mathematical conclusion, but marks for the national examinations belong to their individual subjects. Here EMS is simply shorthand for English, Mathematics and Science working together as a learning method. It does not describe a combined SEC examination paper, and the numerical example should not be mistaken for an official specimen question.
Definition lock: sign, magnitude, mean, variation and model
A signed measurement reports both its numerical value and a chosen reference convention. In a voltage trace, polarity can be positive or negative relative to the connected sensor terminals. Magnitude is the absolute value, written |x|, and it is never negative. The difference between +19 mV and -18 mV as signed values is 37 mV, while the difference between their magnitudes is only 1 mV. Either calculation can be correct, but only one may answer the question asked. Words such as ‘larger’ require you to say whether you mean algebraic value or size of a physical pulse.
The mean is the sum divided by the number of readings. The median is the central value after sorting. The range is maximum minus minimum. These are not interchangeable summaries. The mean responds to every measurement; the median can be less disturbed by an unusual extreme; the range reports spread but does not describe all variation within the set. A small range among repeats may support consistency, yet it does not prove the instrument was accurate. If every measurement was biased by the same calibration error, the readings could agree closely and still be wrong.
A ratio compares two quantities of the same kind after putting them into compatible units. A rate compares change with another quantity, such as millivolts per centimetre per second in a fitted relationship. A residual is the observed value minus the model’s predicted value. A domain is the set of inputs for which a statement is intended. Extrapolation predicts outside the measured interval; interpolation estimates within it. Each term marks a different claim, and a high-quality examination answer knows which one is being made.
Work through the three-speed dataset
The fictional experiment keeps the same coil, magnet, connection convention, sensor and travel path. A student records the peak magnitude of the induced voltage during comparable passes at three nominal speeds. At 5 cm/s, the readings are 9, 10 and 11 mV. At 10 cm/s, the readings are 19, 20 and 21 mV. At 15 cm/s, the readings are 28, 29 and 30 mV. These values form a small example dataset, not measured national examination data. Because the question concerns magnitude, the table lists nonnegative peak sizes rather than signed polarity.
At 5 cm/s, mean = (9 + 10 + 11) / 3 = 10 mV, median = 10 mV and range = 11 – 9 = 2 mV. At 10 cm/s, mean = (19 + 20 + 21) / 3 = 20 mV, median = 20 mV and range = 2 mV. At 15 cm/s, mean = (28 + 29 + 30) / 3 = 29 mV, median = 29 mV and range = 2 mV. The means increase with the nominal speed. Similar ranges indicate comparable numerical spread in these constructed groups, though the relative spread is smaller at the higher means.
From 5 to 10 cm/s the speed doubles and the mean magnitude also doubles, from 10 to 20 mV. From 5 to 15 cm/s speed triples, but the mean magnitude rises from 10 to 29 mV, a factor of 2.9 rather than exactly three. The ratio of mean voltage to speed is 2.0, 2.0 and about 1.93 in the chosen units. This is close to a proportional pattern in the limited dataset, but it is not enough evidence to assert exact proportionality for every speed or apparatus.
Notice what the experimenter did not measure. The data do not report the magnet’s strength, coil resistance, exact peak timing, sensor uncertainty or values above 15 cm/s. The ratios cannot answer those missing questions. A mathematically defensible conclusion says ‘the observed mean peak magnitudes rose approximately linearly over the tested interval, under the stated controls’. A claim that the device will always produce a particular voltage at a much faster speed is a model prediction requiring a separate test.
Build a line carefully, then test it
Plot speed on the horizontal axis and mean peak magnitude on the vertical axis. Name both axes and units: ‘speed (cm/s)’ and ‘mean peak voltage magnitude (mV)’. The three plotted points are (5, 10), (10, 20) and (15, 29). For a simple provisional line using the two endpoint means, the gradient is (29 – 10) / (15 – 5) = 1.9 mV per (cm/s). Substitute (5, 10) into y = 1.9x + c to obtain c = 0.5 mV. Thus y = 1.9x + 0.5 is an illustrative endpoint-based model, not an experimentally established law.
The model predicts 1.9(10) + 0.5 = 19.5 mV at 10 cm/s, whereas the observed mean is 20 mV. The residual, observed minus predicted, is +0.5 mV. For each endpoint the residual is zero by construction. That does not make the two endpoints exceptionally accurate; it only reflects how the line was selected. A true line of best fit is judged using all the data and a stated method. Do not describe an endpoint line as a least-squares result unless least squares was actually used.
If a problem asks for an estimated speed needed to reach 24 mV under this model, solve 1.9x + 0.5 = 24 to obtain x = 23.5 / 1.9 ≈ 12.4 cm/s to three significant figures. This estimate lies inside the tested 5–15 cm/s interval. It remains a model estimate, not a guaranteed physical threshold. The actual sensor response at 12.4 cm/s might vary because of trial scatter, measurement uncertainty and the fact that the fitted relation was approximate.
An estimate at 50 cm/s would be extrapolation far beyond these observations. It may be tempting to insert 50 into the same formula and report 95.5 mV, but the arithmetic is not the main issue. The physical arrangement, magnet motion and sensor performance could behave differently at higher speed. An exam solution may require you to calculate an extrapolated number; when interpreting its reliability, state that it is outside the measured domain and should be tested rather than treated as certain.
Signed readings require a second kind of comparison
Now imagine a forward pass produces +19, +20 and +21 mV at a fixed nominal speed. A reverse pass produces -18, -20 and -19 mV with the same labelled sensor connections. The mean signed forward reading is +20 mV. The mean signed reverse reading is -19 mV. The difference of signed means is 39 mV, which quantifies their separation on the signed scale. It does not mean the forward pulse has 39 mV more magnitude.
The mean of the reverse absolute magnitudes is (18 + 20 + 19) / 3 = 19 mV, versus 20 mV for the forward group. Their difference in average magnitude is 1 mV, or 5% of the forward group’s mean. Even that figure does not establish that withdrawing a magnet is intrinsically less effective than inserting one. The hand motion may have differed; there are only three trials; and we have not quantified the sensor’s uncertainty. The sign contrast is clear in the fictional record, while the size contrast deserves caution.
If six signed readings, +19, +20, +21, -18, -20 and -19, are averaged together, the combined signed mean is 0.5 mV. Calling this ‘the average pulse size’ would be misleading. Positive and negative effects have nearly cancelled in the arithmetic, even though every pulse had a magnitude around 18–21 mV. To answer a magnitude question, first take absolute values, then compute the mean. To answer a net signed-balance question, the signed mean may be appropriate. The question’s wording determines the valid summary.
Percentage, uncertainty and fair comparisons
A change from 10 to 20 mV is an increase of 10 mV, or 100% relative to the original 10 mV. A change from 20 to 29 mV is an increase of 9 mV, or 45% relative to the original 20 mV. The percentage denominators are different. If you simply say ‘the second increase is smaller’, the reader may not know whether you mean absolute change or percentage change. State both when the distinction matters. A table can show equal absolute increases and unequal percentages, or the reverse.
Suppose a sensor is described as having a resolution of 1 mV. This tells you about the smallest displayed increment, not automatically its full measurement uncertainty. Do not turn resolution into a guaranteed ±0.5 mV accuracy claim without a stated instrument specification or an exam assumption. Other sources include variation in movement speed, alignment, starting position and how the logger determines a peak. Identify the source of the uncertainty before proposing a numerical interval or improvement.
A fair comparison controls the measurement context. If one group uses 60 turns and another uses 120 turns, but both change coil diameter, the observed difference cannot be assigned to turns alone. If the peak is recorded from different sensor ranges, first check comparability. If one speed was calculated from a 10 cm path and another from an 8 cm path, verify the rates rather than copying labels. The words ‘same setup’ should mean enough details to support the mathematical inference.
In an examination, an odd reading is not a licence to erase it. Suppose one trial at 10 cm/s is 31 mV rather than 21 mV. The new set is 19, 20, 31, with mean 23.3 mV to three significant figures, median 20 mV and range 12 mV. The mean has shifted sharply while the median stays 20. The reading could be a transcription mistake, a different magnet movement or a genuine observation. Check the record, repeat with controlled conditions and report how the conclusion changes with or without the disputed reading.
A full model answer for a real-world style problem
Consider this original fictional task. A class wants an induction demonstration to exceed an indicative 24 mV peak magnitude without making an unsupported promise. The team has measured the three speed groups above, fits the provisional endpoint model y = 1.9x + 0.5, and may choose from fixed speed settings of 10, 12, 14 or 16 cm/s. Use the model to suggest the smallest setting that meets the numerical target, then explain whether the evidence is sufficient for a guaranteed public claim.
The numerical threshold is x ≥ (24 – 0.5) / 1.9 ≈ 12.368…, so the smallest listed setting satisfying the model is 14 cm/s. At 14 cm/s the model predicts 27.1 mV. The 16 cm/s option also passes the algebraic threshold, but lies just outside the measured speed interval; the 12 cm/s option predicts 23.3 mV and does not pass under this model. The correct numerical selection is 14 cm/s. Do not round the threshold to 12 cm/s and select that setting.
The evidence is not sufficient to guarantee that every 14 cm/s trial will exceed 24 mV. The class has no repeat readings precisely at 14 cm/s, and the model was fitted to only three means. The appropriate recommendation is to test 14 cm/s several times, keep the setup stable and inspect the lowest readings or another suitable reliability measure. The mark-winning answer distinguishes ‘smallest setting according to the model’ from ‘verified setting in actual operation’. This is the difference between doing algebra and making an informed decision.
Twenty-five worked checks for stronger reasoning
1. Positive and negative voltages
A trace shows +14 mV then -17 mV. The greater signed number is +14 because 14 is greater than -17. The greater magnitude is the negative pulse, because |-17| = 17 > 14. A clear response names which property it compares. If an examination asks for polarity, use the sign and the sensor convention. If it asks for strength of the peak, compare absolute values. This single distinction prevents a surprising number of errors in cross-subject numerical questions.
2. Find the mean of three readings
For 7, 8 and 9 mV, the mean is (7 + 8 + 9) / 3 = 8 mV. The arithmetic is simple, but an answer should still identify the measured quantity and unit. If these values are pulse magnitudes, the mean is a mean magnitude. If they are signed readings, you are averaging signed values. The label controls what the result means. A mean is not the maximum and does not prove that every trial will yield exactly 8 mV.
3. What the median protects
Compare 10, 11 and 30 mV. Their mean is 17 mV, while their median is 11 mV. Neither result is wrong. The extreme 30 affects the mean much more. That alone does not prove 30 is a mistake. Inspect the experiment record and repeat a comparable run before deciding whether it is an anomalous observation. A mathematical summary should not be used to conceal inconvenient evidence; use mean, median and range to understand it.
4. A range without an axis label
A graph uses points marked 5, 10 and 15 on the horizontal axis but never says whether they are seconds, centimetres or centimetres per second. The plotted pattern may look convincing, but the interpretation is incomplete. An examination graph should have the measured variable and the unit on each relevant axis. The horizontal axis here is speed in cm/s; the vertical axis is mean peak voltage magnitude in mV. Context is part of mathematics, not an optional caption.
5. Percentage increase from a baseline
A peak rises from 12 mV to 18 mV. The increase is 6 mV, and the percentage increase relative to 12 mV is 6/12 × 100% = 50%. Dividing by 18 instead gives 33.3%, which answers a different question. Identify the original value whenever the word ‘increase’ is used. If the task compares two methods, explain which method is the baseline and why. A percentage without its denominator is often a misleading statistic.
6. Comparing ratios
Method A has a mean 12 mV at 6 cm/s. Method B has a mean 24 mV at 12 cm/s. Both have 2 mV per (cm/s) as a simple quotient. The equal ratios may suggest approximate proportionality under those chosen conditions, but two points do not establish an exact physical rule. The phrase ‘mean voltage divided by speed’ describes the computation; the word ‘gradient’ requires a defined graph and a comparison of changes between points. Keep the mathematical terms distinct.
7. The gradient of two points
The points (5,10) and (15,29) give gradient (29-10)/(15-5) = 1.9. In context it is about 1.9 mV change in modelled peak magnitude for each additional cm/s of speed within the interval. Do not report merely ‘1.9 volts’. The units of gradient are vertical units divided by horizontal units. A gradient describes rate of change in a model. Whether the model continues to hold physically is another question.
8. The intercept is not automatically physical
The endpoint model y = 1.9x + 0.5 predicts 0.5 mV at x = 0. Yet the physics experiment suggests no sustained induction in this stationary arrangement, and a fitted intercept can reflect scatter or model choice. Do not elevate 0.5 mV into a discovery that a motionless magnet generates voltage. The intercept is a mathematical property of the provisional line. Interpret it with knowledge of the measurement setup and its range.
9. A residual with a sign
If the observed magnitude is 20 mV while the model predicts 19.5 mV, residual = observed – predicted = +0.5 mV. If the observed value is 18 mV instead, the residual is -1.5 mV. A positive residual means the measurement lies above the model prediction, not that the sensor polarity is positive. The two different uses of signs must not be confused. Label the residual and the original signed voltage separately if both appear.
10. Interpolate instead of guessing
Suppose the provisional model is y = 1.9x + 0.5 and x = 12 cm/s. Substitution gives y = 23.3 mV. Because x falls inside the tested 5–15 cm/s interval, this is interpolation. It is an estimate, not a measurement. If the exam prompt says ‘use the model’, show the substitution. If it asks whether the experiment proves the result, explain that no 12 cm/s trial has yet been performed in the dataset.
11. An integer-choice constraint
A demonstration can choose speeds only from 8, 10, 12, 14 and 16 cm/s. A model threshold gives x ≥ 12.37 cm/s. The smallest feasible listed speed is 14 cm/s, not 12.37 or 13. The answer must obey both the algebraic inequality and the permitted choices. For an equipment decision, add that 14 cm/s is an unverified model choice until trials support it. Mathematically feasible is not always experimentally established.
12. Difference between two means
Forward magnitude readings have mean 20 mV; reverse magnitude readings have mean 19 mV. The absolute mean difference is 1 mV and the difference relative to forward is 5%. A student who subtracts the signed means +20 and -19 obtains 39 mV, which is the separation of their signed positions, not their relative pulse sizes. State the quantity first. The problem determines whether to calculate the magnitude difference or the signed difference.
13. Why averaging signed pulses fails
Combine +20 and -20 mV into a mean of zero. This correctly describes the signed average, but a report saying ‘the apparatus produced no voltage’ would be false. There were two nonzero pulses with opposite polarities. In data involving directions, cancellation is not the same as absence. Report signed mean if net polarity balance is relevant, or mean absolute magnitude if the task concerns the size of the pulses. Choose the summary that retains the information the reader needs.
14. Three significant figures
An inverse model yields x = 23.5 / 1.9 = 12.368421…. To three significant figures, report 12.4 cm/s. If the choice is constrained to integer speeds, do not use 12.4 as though it were an allowed setting. If the question requires two decimal places instead, follow its instruction. Show enough intermediate accuracy to avoid changing the decision by premature rounding. The result’s numerical precision must fit both the syllabus convention and the problem.
15. Instrument resolution
If a sensor displays to the nearest 1 mV, recorded values of 19, 20 and 21 mV should not automatically be treated as exact real numbers. Nor should a student claim a calibrated uncertainty of ±0.5 mV without an instrument specification. Resolution is about display steps, while accuracy and uncertainty depend on additional information. In a written response, name the missing specification and explain why it matters if the question asks about reliability.
16. A highly correlated pattern
Three means rise as speed rises. This is consistent with a positive association in this small dataset. It is not a universal guarantee that speed alone caused every difference or that the same model holds indefinitely. Controlled experimental design supports a causal interpretation within its tested setting, but unmeasured factors can remain. A student should combine mathematical evidence with the science method rather than treating a straight-looking graph as conclusive by itself.
17. The anomaly question
The readings at one speed are 19, 20 and 31 mV. The mean is 70/3 ≈ 23.3 mV; the median is 20 mV; the range is 12 mV. Instead of deleting 31 and announcing the mean is 19.5, investigate why that run differed. Recheck the recorded speed, path, sensor range and raw log. If an error is confirmed, explain it and retain an audit trail. If not, present the variation honestly and consider more repeats.
18. A false efficiency comparison
A learner calls the setup with the highest peak magnitude ‘most efficient’. Efficiency compares useful energy output with input energy, not just peak voltage. The table contains no input work or delivered energy. The mathematically sound claim is that one tested condition had the highest recorded peak magnitude. The physical efficiency question needs different evidence. This is a valuable reminder that correct arithmetic cannot rescue a claim about a quantity that was never measured.
19. Exact versus approximate
The results (5,10), (10,20), (15,29) are not exactly proportional because the ratios y/x are not identical. They may be described as approximately proportional over the tested interval. You can quantify the departure or inspect the residual of a provisional line. Avoid hiding the 29 by rounding it to 30 merely to make a prettier pattern. The model must serve the data rather than the data being adjusted to serve the model.
20. A revised conclusion
An original conclusion says ‘our coil always doubles its voltage when speed doubles’. The revised version says: ‘For the three tested speeds and the unchanged apparatus, the mean peak magnitude increased from 10 to 20 to 29 mV and followed an approximately linear pattern.’ The statement gives the observable evidence, the range and the qualification. It can be defended without speculating about every magnet, coil or speed. In applied Mathematics, precision of language protects the value of the calculation.
21. Reading a poorly scaled graph
A graph starts the vertical axis at 18 mV rather than zero, making a change from 19 to 21 mV appear enormous. Such scales are not automatically forbidden, but they may distort a casual impression of relative differences. Check the axis baseline, intervals and labels before using visual steepness as evidence. If the graph is intended for a general audience, report the actual numerical change and consider whether the scale communicates proportionate meaning.
22. Matching a line to a domain
A straight-line model fitted for 5–15 cm/s is used to predict -5 cm/s. Negative speed is not meaningful if the variable is speed as a nonnegative magnitude; reversing direction belongs in a separate signed-velocity description. The model’s domain and variable definition matter. A numerical substitution may be possible algebraically but physically irrelevant. A good exam response first translates the question into a meaningful variable before solving.
23. Rate from distance and time
A magnet travels 10 cm in 2 seconds over a prescribed path. Average speed is 10/2 = 5 cm/s. If it travels 10 cm in 1 second, the average speed is 10 cm/s. These rates can serve as controlled categories, but the movement may not have been uniform throughout the pass. A peak voltage is sensitive to the rate of change near a particular moment, so average speed over the whole path may be an imperfect proxy. State that limitation when evaluating the model.
24. What repeating a trial does
Three similar measurements can reveal variation in observed peaks and make the numerical summary more informative. They do not automatically guarantee that the sensor calibration, path or starting distance is correct. Repeated measurements improve evidence about repeatability under the trial protocol, while accuracy requires appropriate checks against known standards or clearly defined physical quantities. The distinction prevents vague statements that ‘three repeats prove accuracy’.
25. How to present final working
Write the formula, substitute numbers with units, calculate without unnecessary early rounding, and end with a context sentence. For the threshold example: 1.9x + 0.5 ≥ 24, so x ≥ 12.37… cm/s; among the given settings, choose 14 cm/s. Then say the choice is model-based and should be verified by repeat trials. This compact structure is often stronger than several paragraphs that never identify a mathematical constraint or state the final decision.
Practice questions and answer checks
Practice A: the magnitudes 8, 10, 12 and 10 mV are observed at one speed. Find mean, median and range. Answer: mean 10 mV, median 10 mV and range 4 mV. Practice B: a pulse is -25 mV and another is +22 mV. State which has the larger algebraic value and which has the larger magnitude. Answer: +22 mV is algebraically larger, but the -25 mV pulse has a magnitude of 25 mV and is larger in absolute size.
Practice C: a model predicts y = 1.9x + 0.5. Find y at x = 11 cm/s. Answer: 21.4 mV. Practice D: find x when y = 24 mV. Answer: 23.5/1.9 ≈ 12.4 cm/s. Practice E: explain why an unmeasured prediction at x = 40 cm/s is weaker than one at 11 cm/s. Answer: 40 is outside the tested interval and requires extrapolation; the model’s validity there is unknown.
Practice F: a table shows 12, 12 and 36 mV and a student discards 36 ‘to make the results consistent’. Explain the correct approach. Answer: keep the original observation, check transcription, instrument and movement, repeat under controlled conditions, and only exclude an error with an explicit justification. Practice G: describe the best conclusion for the three-speed data. Answer: mean peak magnitudes increased approximately linearly from 10 to 29 mV over the tested speeds under the same setup; more measurements are needed for wider claims.
Where this guide belongs
This volume builds on the Physics induction practical in Vol 0500 and precedes the English and EMS transfer work in this sequence. Students who need a foundation in graph reading can revisit Vol 0465, graph rates and unit checks, or Vol 0481, repeat-trial statistics. The G3 SEC EMS learner guide is the larger route.
Before making a final decision, check what the numbers actually represent, whether the comparison was fair, whether the chosen model is justified and whether the proposed sentence says more than the data allow. This habit began at primary level with careful word-problem interpretation; the PSLE Mathematics word-problem bridge remains useful when a G3 task feels unfamiliar. The algebra is more advanced, but the underlying question remains simple: what is known, what is being asked, and what reasoning connects the two?
