G2 Science K223, K224 and K225 experimental reasoning becomes more reliable when learners understand calibration: using known standards or reference conditions to establish how an instrument signal, test response or observable measurement maps onto an unknown quantity.
This one-hundred-and-sixty-seventh Learner’s Guide develops calibration-curve reasoning. The goal is not advanced laboratory statistics. The goal is to understand why known references matter, why interpolation is safer than unsupported extrapolation, and why a calibrated measurement can still fail if the sample, timing or instrument changes.
Mechanism: standards create the measurement mapping
A detector often gives a signal, not the target quantity directly. Calibration links known input values to measured outputs. Once the relationship is established over a valid range, an unknown signal can be mapped back to an estimated quantity.
Diagnosis
When a reading looks suspicious, separate four possibilities: the sample truly differs, the instrument mapping is wrong, the reference standard is wrong, or the sample falls outside the calibrated range. Diagnose with standards and controls before blaming the sample.
Smallest repair
Repeat the smallest reference capable of testing the suspected failure: blank for offset/background, known positive for sensitivity, a second standard for slope, or a post-calibration check for drift. Recalibrate only as much as the diagnosis requires.
1. known zero
Calibration principle: Use a blank or zero-input condition to define the baseline reading.
Control: A nonzero blank reveals background signal or offset.
For practice, sketch a simple standard-response graph and place an unknown signal on it. State whether the unknown is interpolated, extrapolated, below detection, or near saturation, then explain what conclusion is justified.
For transfer, change the instrument, units, background, time or response shape. The learner should rebuild the calibration logic rather than reuse one conversion factor blindly.
2. known low standard
Calibration principle: Choose a reference near the lower useful range.
Control: It helps anchor interpolation for small unknowns.
For practice, sketch a simple standard-response graph and place an unknown signal on it. State whether the unknown is interpolated, extrapolated, below detection, or near saturation, then explain what conclusion is justified.
For transfer, change the instrument, units, background, time or response shape. The learner should rebuild the calibration logic rather than reuse one conversion factor blindly.
3. known high standard
Calibration principle: Choose a reference near the upper useful range.
Control: Unknowns above it require extrapolation or dilution/new range.
For practice, sketch a simple standard-response graph and place an unknown signal on it. State whether the unknown is interpolated, extrapolated, below detection, or near saturation, then explain what conclusion is justified.
For transfer, change the instrument, units, background, time or response shape. The learner should rebuild the calibration logic rather than reuse one conversion factor blindly.
4. two-point calibration
Calibration principle: Two standards define a straight-line mapping only if linear response is justified.
Control: Check a third standard when possible.
For practice, sketch a simple standard-response graph and place an unknown signal on it. State whether the unknown is interpolated, extrapolated, below detection, or near saturation, then explain what conclusion is justified.
For transfer, change the instrument, units, background, time or response shape. The learner should rebuild the calibration logic rather than reuse one conversion factor blindly.
5. multi-point calibration
Calibration principle: Several standards reveal whether response is linear, curved or saturating.
Control: Use the shape rather than forcing a straight line.
For practice, sketch a simple standard-response graph and place an unknown signal on it. State whether the unknown is interpolated, extrapolated, below detection, or near saturation, then explain what conclusion is justified.
For transfer, change the instrument, units, background, time or response shape. The learner should rebuild the calibration logic rather than reuse one conversion factor blindly.
6. blank subtraction
Calibration principle: Subtract background only when the blank represents the same background process as the samples.
Control: A mismatched blank can create bias.
For practice, sketch a simple standard-response graph and place an unknown signal on it. State whether the unknown is interpolated, extrapolated, below detection, or near saturation, then explain what conclusion is justified.
For transfer, change the instrument, units, background, time or response shape. The learner should rebuild the calibration logic rather than reuse one conversion factor blindly.
7. sensor zero
Calibration principle: Set or record instrument zero before measurement.
Control: Zero drift later can invalidate the original correction.
For practice, sketch a simple standard-response graph and place an unknown signal on it. State whether the unknown is interpolated, extrapolated, below detection, or near saturation, then explain what conclusion is justified.
For transfer, change the instrument, units, background, time or response shape. The learner should rebuild the calibration logic rather than reuse one conversion factor blindly.
8. known reference sample
Calibration principle: Measure a sample with established value.
Control: It tests both instrument response and procedure.
For practice, sketch a simple standard-response graph and place an unknown signal on it. State whether the unknown is interpolated, extrapolated, below detection, or near saturation, then explain what conclusion is justified.
For transfer, change the instrument, units, background, time or response shape. The learner should rebuild the calibration logic rather than reuse one conversion factor blindly.
9. positive control
Calibration principle: A known-positive response shows the method can detect the target under current conditions.
Control: Failure weakens all negative sample conclusions.
For practice, sketch a simple standard-response graph and place an unknown signal on it. State whether the unknown is interpolated, extrapolated, below detection, or near saturation, then explain what conclusion is justified.
For transfer, change the instrument, units, background, time or response shape. The learner should rebuild the calibration logic rather than reuse one conversion factor blindly.
10. negative control
Calibration principle: A known-negative response reveals false positive/background behaviour.
Control: Unexpected signal indicates specificity or contamination issues.
For practice, sketch a simple standard-response graph and place an unknown signal on it. State whether the unknown is interpolated, extrapolated, below detection, or near saturation, then explain what conclusion is justified.
For transfer, change the instrument, units, background, time or response shape. The learner should rebuild the calibration logic rather than reuse one conversion factor blindly.
11. calibration range
Calibration principle: Keep unknowns inside the range covered by standards when possible.
Control: Interpolation is safer than extrapolation.
For practice, sketch a simple standard-response graph and place an unknown signal on it. State whether the unknown is interpolated, extrapolated, below detection, or near saturation, then explain what conclusion is justified.
For transfer, change the instrument, units, background, time or response shape. The learner should rebuild the calibration logic rather than reuse one conversion factor blindly.
12. interpolation
Calibration principle: Estimate an unknown between nearby standards.
Control: Uncertainty depends on spacing, noise and curve shape.
For practice, sketch a simple standard-response graph and place an unknown signal on it. State whether the unknown is interpolated, extrapolated, below detection, or near saturation, then explain what conclusion is justified.
For transfer, change the instrument, units, background, time or response shape. The learner should rebuild the calibration logic rather than reuse one conversion factor blindly.
13. extrapolation
Calibration principle: Predict beyond the standards only with explicit caution.
Control: The response relationship may change outside the calibrated range.
For practice, sketch a simple standard-response graph and place an unknown signal on it. State whether the unknown is interpolated, extrapolated, below detection, or near saturation, then explain what conclusion is justified.
For transfer, change the instrument, units, background, time or response shape. The learner should rebuild the calibration logic rather than reuse one conversion factor blindly.
14. linear response
Calibration principle: Signal changes proportionally or approximately linearly with target over a stated range.
Control: Do not assume linearity from two points alone.
For practice, sketch a simple standard-response graph and place an unknown signal on it. State whether the unknown is interpolated, extrapolated, below detection, or near saturation, then explain what conclusion is justified.
For transfer, change the instrument, units, background, time or response shape. The learner should rebuild the calibration logic rather than reuse one conversion factor blindly.
15. nonlinear response
Calibration principle: Curve bends across the range.
Control: Use the calibrated shape rather than one constant conversion factor.
For practice, sketch a simple standard-response graph and place an unknown signal on it. State whether the unknown is interpolated, extrapolated, below detection, or near saturation, then explain what conclusion is justified.
For transfer, change the instrument, units, background, time or response shape. The learner should rebuild the calibration logic rather than reuse one conversion factor blindly.
16. saturation
Calibration principle: Signal stops increasing meaningfully at high input.
Control: High unknowns can become indistinguishable.
For practice, sketch a simple standard-response graph and place an unknown signal on it. State whether the unknown is interpolated, extrapolated, below detection, or near saturation, then explain what conclusion is justified.
For transfer, change the instrument, units, background, time or response shape. The learner should rebuild the calibration logic rather than reuse one conversion factor blindly.
17. detection floor
Calibration principle: Small inputs produce signal comparable to background.
Control: Do not claim precise low values below useful resolution.
For practice, sketch a simple standard-response graph and place an unknown signal on it. State whether the unknown is interpolated, extrapolated, below detection, or near saturation, then explain what conclusion is justified.
For transfer, change the instrument, units, background, time or response shape. The learner should rebuild the calibration logic rather than reuse one conversion factor blindly.
18. resolution
Calibration principle: Instrument changes in steps or finite increments.
Control: Report precision consistent with readable increments.
For practice, sketch a simple standard-response graph and place an unknown signal on it. State whether the unknown is interpolated, extrapolated, below detection, or near saturation, then explain what conclusion is justified.
For transfer, change the instrument, units, background, time or response shape. The learner should rebuild the calibration logic rather than reuse one conversion factor blindly.
19. repeat standards
Calibration principle: Repeated reference readings reveal short-term variation.
Control: Spread matters for how precisely unknowns can be read.
For practice, sketch a simple standard-response graph and place an unknown signal on it. State whether the unknown is interpolated, extrapolated, below detection, or near saturation, then explain what conclusion is justified.
For transfer, change the instrument, units, background, time or response shape. The learner should rebuild the calibration logic rather than reuse one conversion factor blindly.
20. repeat unknown
Calibration principle: Multiple sample readings reveal repeatability but not calibration validity by themselves.
Control: Standards are still needed.
For practice, sketch a simple standard-response graph and place an unknown signal on it. State whether the unknown is interpolated, extrapolated, below detection, or near saturation, then explain what conclusion is justified.
For transfer, change the instrument, units, background, time or response shape. The learner should rebuild the calibration logic rather than reuse one conversion factor blindly.
21. same procedure
Calibration principle: Standards and unknowns should be prepared/measured comparably.
Control: Different handling can break the calibration relation.
For practice, sketch a simple standard-response graph and place an unknown signal on it. State whether the unknown is interpolated, extrapolated, below detection, or near saturation, then explain what conclusion is justified.
For transfer, change the instrument, units, background, time or response shape. The learner should rebuild the calibration logic rather than reuse one conversion factor blindly.
22. same units
Calibration principle: Reference values and unknown result must share compatible units.
Control: Conversion errors can masquerade as calibration errors.
For practice, sketch a simple standard-response graph and place an unknown signal on it. State whether the unknown is interpolated, extrapolated, below detection, or near saturation, then explain what conclusion is justified.
For transfer, change the instrument, units, background, time or response shape. The learner should rebuild the calibration logic rather than reuse one conversion factor blindly.
23. same matrix/context
Calibration principle: Background composition can affect signal.
Control: A standard in a very different environment may not transfer.
For practice, sketch a simple standard-response graph and place an unknown signal on it. State whether the unknown is interpolated, extrapolated, below detection, or near saturation, then explain what conclusion is justified.
For transfer, change the instrument, units, background, time or response shape. The learner should rebuild the calibration logic rather than reuse one conversion factor blindly.
24. same temperature
Calibration principle: Instrument or reaction response can depend on temperature.
Control: Control or record relevant conditions.
For practice, sketch a simple standard-response graph and place an unknown signal on it. State whether the unknown is interpolated, extrapolated, below detection, or near saturation, then explain what conclusion is justified.
For transfer, change the instrument, units, background, time or response shape. The learner should rebuild the calibration logic rather than reuse one conversion factor blindly.
25. same timing
Calibration principle: Read standards and unknowns at comparable times after preparation.
Control: Response lag can otherwise distort mapping.
For practice, sketch a simple standard-response graph and place an unknown signal on it. State whether the unknown is interpolated, extrapolated, below detection, or near saturation, then explain what conclusion is justified.
For transfer, change the instrument, units, background, time or response shape. The learner should rebuild the calibration logic rather than reuse one conversion factor blindly.
26. same apparatus
Calibration principle: Different instruments can have different offsets/gains.
Control: Each may require its own calibration.
For practice, sketch a simple standard-response graph and place an unknown signal on it. State whether the unknown is interpolated, extrapolated, below detection, or near saturation, then explain what conclusion is justified.
For transfer, change the instrument, units, background, time or response shape. The learner should rebuild the calibration logic rather than reuse one conversion factor blindly.
27. calibration before experiment
Calibration principle: Establish mapping before unknown measurements.
Control: But this does not guarantee stability later.
For practice, sketch a simple standard-response graph and place an unknown signal on it. State whether the unknown is interpolated, extrapolated, below detection, or near saturation, then explain what conclusion is justified.
For transfer, change the instrument, units, background, time or response shape. The learner should rebuild the calibration logic rather than reuse one conversion factor blindly.
28. calibration after experiment
Calibration principle: A post-check can reveal drift during the session.
Control: Agreement before/after strengthens confidence.
For practice, sketch a simple standard-response graph and place an unknown signal on it. State whether the unknown is interpolated, extrapolated, below detection, or near saturation, then explain what conclusion is justified.
For transfer, change the instrument, units, background, time or response shape. The learner should rebuild the calibration logic rather than reuse one conversion factor blindly.
29. mid-session calibration
Calibration principle: Periodic standards can locate when drift begins.
Control: Useful in long runs.
For practice, sketch a simple standard-response graph and place an unknown signal on it. State whether the unknown is interpolated, extrapolated, below detection, or near saturation, then explain what conclusion is justified.
For transfer, change the instrument, units, background, time or response shape. The learner should rebuild the calibration logic rather than reuse one conversion factor blindly.
30. drift
Calibration principle: Reference reading changes with time.
Control: Recalibration or time-matched references may be needed.
For practice, sketch a simple standard-response graph and place an unknown signal on it. State whether the unknown is interpolated, extrapolated, below detection, or near saturation, then explain what conclusion is justified.
For transfer, change the instrument, units, background, time or response shape. The learner should rebuild the calibration logic rather than reuse one conversion factor blindly.
31. warm-up
Calibration principle: Early instrument response may be unstable.
Control: Allow stabilisation before trusting calibration.
For practice, sketch a simple standard-response graph and place an unknown signal on it. State whether the unknown is interpolated, extrapolated, below detection, or near saturation, then explain what conclusion is justified.
For transfer, change the instrument, units, background, time or response shape. The learner should rebuild the calibration logic rather than reuse one conversion factor blindly.
32. carryover
Calibration principle: Previous high sample affects next reading.
Control: Rinse/reset and include sequence controls.
For practice, sketch a simple standard-response graph and place an unknown signal on it. State whether the unknown is interpolated, extrapolated, below detection, or near saturation, then explain what conclusion is justified.
For transfer, change the instrument, units, background, time or response shape. The learner should rebuild the calibration logic rather than reuse one conversion factor blindly.
33. contamination
Calibration principle: Standards or unknowns are altered by foreign material.
Control: Calibration cannot rescue a contaminated sample.
For practice, sketch a simple standard-response graph and place an unknown signal on it. State whether the unknown is interpolated, extrapolated, below detection, or near saturation, then explain what conclusion is justified.
For transfer, change the instrument, units, background, time or response shape. The learner should rebuild the calibration logic rather than reuse one conversion factor blindly.
34. standard preparation error
Calibration principle: Reference value itself is wrong.
Control: All unknowns can then be biased systematically.
For practice, sketch a simple standard-response graph and place an unknown signal on it. State whether the unknown is interpolated, extrapolated, below detection, or near saturation, then explain what conclusion is justified.
For transfer, change the instrument, units, background, time or response shape. The learner should rebuild the calibration logic rather than reuse one conversion factor blindly.
35. label swap
Calibration principle: Standards are assigned to wrong concentrations/values.
Control: Curve shape may look impossible or reversed.
For practice, sketch a simple standard-response graph and place an unknown signal on it. State whether the unknown is interpolated, extrapolated, below detection, or near saturation, then explain what conclusion is justified.
For transfer, change the instrument, units, background, time or response shape. The learner should rebuild the calibration logic rather than reuse one conversion factor blindly.
36. axis swap
Calibration principle: Signal and known value are plotted on wrong axes relative to intended read-off.
Control: Keep input/reference and output/signal ownership explicit.
For practice, sketch a simple standard-response graph and place an unknown signal on it. State whether the unknown is interpolated, extrapolated, below detection, or near saturation, then explain what conclusion is justified.
For transfer, change the instrument, units, background, time or response shape. The learner should rebuild the calibration logic rather than reuse one conversion factor blindly.
37. wrong scale
Calibration principle: Graph spacing or units are misread.
Control: Check tick intervals before reading unknown.
For practice, sketch a simple standard-response graph and place an unknown signal on it. State whether the unknown is interpolated, extrapolated, below detection, or near saturation, then explain what conclusion is justified.
For transfer, change the instrument, units, background, time or response shape. The learner should rebuild the calibration logic rather than reuse one conversion factor blindly.
38. origin truncation
Calibration principle: Graph may not start at zero.
Control: Do not infer proportionality from visual appearance alone.
For practice, sketch a simple standard-response graph and place an unknown signal on it. State whether the unknown is interpolated, extrapolated, below detection, or near saturation, then explain what conclusion is justified.
For transfer, change the instrument, units, background, time or response shape. The learner should rebuild the calibration logic rather than reuse one conversion factor blindly.
39. best-fit relation
Calibration principle: Standards may scatter around a trend.
Control: Use the intended school-level trend without pretending every point is exact.
For practice, sketch a simple standard-response graph and place an unknown signal on it. State whether the unknown is interpolated, extrapolated, below detection, or near saturation, then explain what conclusion is justified.
For transfer, change the instrument, units, background, time or response shape. The learner should rebuild the calibration logic rather than reuse one conversion factor blindly.
40. outlier standard
Calibration principle: One reference point conflicts with the rest.
Control: Repeat or inspect preparation before deleting it.
For practice, sketch a simple standard-response graph and place an unknown signal on it. State whether the unknown is interpolated, extrapolated, below detection, or near saturation, then explain what conclusion is justified.
For transfer, change the instrument, units, background, time or response shape. The learner should rebuild the calibration logic rather than reuse one conversion factor blindly.
41. unknown between standards
Calibration principle: Use neighbouring points or fitted relation.
Control: Do not use distant anchors unnecessarily.
For practice, sketch a simple standard-response graph and place an unknown signal on it. State whether the unknown is interpolated, extrapolated, below detection, or near saturation, then explain what conclusion is justified.
For transfer, change the instrument, units, background, time or response shape. The learner should rebuild the calibration logic rather than reuse one conversion factor blindly.
42. unknown near boundary
Calibration principle: Reading uncertainty can push it inside/outside range.
Control: State caution near limits.
For practice, sketch a simple standard-response graph and place an unknown signal on it. State whether the unknown is interpolated, extrapolated, below detection, or near saturation, then explain what conclusion is justified.
For transfer, change the instrument, units, background, time or response shape. The learner should rebuild the calibration logic rather than reuse one conversion factor blindly.
43. unknown above range
Calibration principle: Dilute/change range or report that calibration does not support precise value.
Control: Avoid blind extrapolation.
For practice, sketch a simple standard-response graph and place an unknown signal on it. State whether the unknown is interpolated, extrapolated, below detection, or near saturation, then explain what conclusion is justified.
For transfer, change the instrument, units, background, time or response shape. The learner should rebuild the calibration logic rather than reuse one conversion factor blindly.
44. unknown below range
Calibration principle: Use a more sensitive method/range if available.
Control: Do not invent precision.
For practice, sketch a simple standard-response graph and place an unknown signal on it. State whether the unknown is interpolated, extrapolated, below detection, or near saturation, then explain what conclusion is justified.
For transfer, change the instrument, units, background, time or response shape. The learner should rebuild the calibration logic rather than reuse one conversion factor blindly.
45. calibration slope
Calibration principle: A steeper response means signal changes more per unit input.
Control: Sensitivity depends on slope and noise together.
For practice, sketch a simple standard-response graph and place an unknown signal on it. State whether the unknown is interpolated, extrapolated, below detection, or near saturation, then explain what conclusion is justified.
For transfer, change the instrument, units, background, time or response shape. The learner should rebuild the calibration logic rather than reuse one conversion factor blindly.
46. calibration intercept
Calibration principle: Nonzero intercept can represent background/offset.
Control: Do not force through zero unless justified.
For practice, sketch a simple standard-response graph and place an unknown signal on it. State whether the unknown is interpolated, extrapolated, below detection, or near saturation, then explain what conclusion is justified.
For transfer, change the instrument, units, background, time or response shape. The learner should rebuild the calibration logic rather than reuse one conversion factor blindly.
47. gain change
Calibration principle: Slope changes over time or between instruments.
Control: Old conversion factor may no longer apply.
For practice, sketch a simple standard-response graph and place an unknown signal on it. State whether the unknown is interpolated, extrapolated, below detection, or near saturation, then explain what conclusion is justified.
For transfer, change the instrument, units, background, time or response shape. The learner should rebuild the calibration logic rather than reuse one conversion factor blindly.
48. offset change
Calibration principle: Whole calibration shifts up/down.
Control: Blank or zero check can detect it.
For practice, sketch a simple standard-response graph and place an unknown signal on it. State whether the unknown is interpolated, extrapolated, below detection, or near saturation, then explain what conclusion is justified.
For transfer, change the instrument, units, background, time or response shape. The learner should rebuild the calibration logic rather than reuse one conversion factor blindly.
49. slope and offset both change
Calibration principle: A single zero correction may be insufficient.
Control: Use multiple standards.
For practice, sketch a simple standard-response graph and place an unknown signal on it. State whether the unknown is interpolated, extrapolated, below detection, or near saturation, then explain what conclusion is justified.
For transfer, change the instrument, units, background, time or response shape. The learner should rebuild the calibration logic rather than reuse one conversion factor blindly.
50. reference hierarchy
Calibration principle: Some standards are more trustworthy than others.
Control: Prefer traceable/defined references given by the task.
For practice, sketch a simple standard-response graph and place an unknown signal on it. State whether the unknown is interpolated, extrapolated, below detection, or near saturation, then explain what conclusion is justified.
For transfer, change the instrument, units, background, time or response shape. The learner should rebuild the calibration logic rather than reuse one conversion factor blindly.
51. calibration and proxy
Calibration principle: Signal is a proxy for target quantity.
Control: Validity depends on the relationship established by standards.
For practice, sketch a simple standard-response graph and place an unknown signal on it. State whether the unknown is interpolated, extrapolated, below detection, or near saturation, then explain what conclusion is justified.
For transfer, change the instrument, units, background, time or response shape. The learner should rebuild the calibration logic rather than reuse one conversion factor blindly.
52. calibration and causality
Calibration principle: Calibration maps signal to quantity; it does not prove what caused that quantity.
Control: Keep measurement and mechanism separate.
For practice, sketch a simple standard-response graph and place an unknown signal on it. State whether the unknown is interpolated, extrapolated, below detection, or near saturation, then explain what conclusion is justified.
For transfer, change the instrument, units, background, time or response shape. The learner should rebuild the calibration logic rather than reuse one conversion factor blindly.
53. calibration and sample representativeness
Calibration principle: A precise calibrated reading from one location may still be unrepresentative of the whole system.
Control: Measurement validity and sampling validity are different.
For practice, sketch a simple standard-response graph and place an unknown signal on it. State whether the unknown is interpolated, extrapolated, below detection, or near saturation, then explain what conclusion is justified.
For transfer, change the instrument, units, background, time or response shape. The learner should rebuild the calibration logic rather than reuse one conversion factor blindly.
54. calibration and temporal cadence
Calibration principle: A calibrated sensor can still miss fast changes if sampled too slowly.
Control: Calibration does not replace timing design.
For practice, sketch a simple standard-response graph and place an unknown signal on it. State whether the unknown is interpolated, extrapolated, below detection, or near saturation, then explain what conclusion is justified.
For transfer, change the instrument, units, background, time or response shape. The learner should rebuild the calibration logic rather than reuse one conversion factor blindly.
55. calibration and spatial gradient
Calibration principle: A calibrated reading can vary by location because the system truly varies.
Control: Do not call every difference instrument error.
For practice, sketch a simple standard-response graph and place an unknown signal on it. State whether the unknown is interpolated, extrapolated, below detection, or near saturation, then explain what conclusion is justified.
For transfer, change the instrument, units, background, time or response shape. The learner should rebuild the calibration logic rather than reuse one conversion factor blindly.
56. calibration and compartment model
Calibration principle: Different compartments may require separate interpretation even with same sensor.
Control: Location ownership matters.
For practice, sketch a simple standard-response graph and place an unknown signal on it. State whether the unknown is interpolated, extrapolated, below detection, or near saturation, then explain what conclusion is justified.
For transfer, change the instrument, units, background, time or response shape. The learner should rebuild the calibration logic rather than reuse one conversion factor blindly.
57. calibration and negative evidence
Calibration principle: No detected signal is stronger when the method is calibrated to detect the expected range.
Control: Calibration supports sensitivity claims.
For practice, sketch a simple standard-response graph and place an unknown signal on it. State whether the unknown is interpolated, extrapolated, below detection, or near saturation, then explain what conclusion is justified.
For transfer, change the instrument, units, background, time or response shape. The learner should rebuild the calibration logic rather than reuse one conversion factor blindly.
58. calibration and positive evidence
Calibration principle: A positive signal should lie within validated response range.
Control: Saturation can make high values ambiguous.
For practice, sketch a simple standard-response graph and place an unknown signal on it. State whether the unknown is interpolated, extrapolated, below detection, or near saturation, then explain what conclusion is justified.
For transfer, change the instrument, units, background, time or response shape. The learner should rebuild the calibration logic rather than reuse one conversion factor blindly.
59. calibration curve final rule
Calibration principle: Use known standards to establish how signal maps to quantity, keep unknowns within the validated range, and recheck the reference whenever the system or instrument may have drifted.
Control: undefined
For practice, sketch a simple standard-response graph and place an unknown signal on it. State whether the unknown is interpolated, extrapolated, below detection, or near saturation, then explain what conclusion is justified.
For transfer, change the instrument, units, background, time or response shape. The learner should rebuild the calibration logic rather than reuse one conversion factor blindly.
Competing explanations
If an unknown reading changes, competing explanations include real sample change, instrument drift, blank shift, gain change, contamination or timing differences. Choose a reference measurement that these explanations predict differently.
Observation versus inference
The instrument signal is observed. The converted concentration, intensity or quantity is inferred through the calibration relationship. The claim about what caused that quantity is a further inference. Keep all three layers separate.
Support versus proof
A good calibration supports accurate measurement within the tested range. It does not prove the instrument will behave identically outside that range, under a different background, or after long drift. Validation remains conditional.
Model limits
School-level calibration curves may be treated as linear in a stated range, but real response can curve or saturate. Follow the evidence supplied by the task and avoid forcing a straight-line model when the standards show otherwise.
Delayed transfer
Return later with an unfamiliar sensor or test where standards are given but the calibration method is not named. Ask the learner to identify the mapping, locate the valid range, read an unknown, diagnose one bad reference and state what cannot be concluded outside the calibrated conditions.
Internal learning links
Use the Science Hub, Vol 0159 Sampling Representativeness, Vol 0163 Temporal Sampling Cadence, Examination Craft and the PSLE Learner’s Guide.
MOE and SEAB current framework
MOE states that Full Subject-Based Banding is fully implemented and that, from 2027, the Singapore-Cambridge Secondary Education Certificate (SEC) replaces the former N- and O-Level examinations, with graduating students sitting subjects at their respective G1, G2 or G3 levels. See the official MOE Full SBB / SEC announcement. For the current 2027 G2 Science combinations K223, K224 and K225 and linked syllabuses, use the official SEAB G2 syllabus directory.
Final rule
A measurement is trustworthy only when the reference mapping is trustworthy. Calibrate with known standards, keep unknowns inside the validated range, distinguish signal from inferred quantity, and recheck the reference whenever drift or background change is plausible.
G2 SEC Learner’s Guide: Open the Vol 0132–0175 hub and subject index.
