PSLE-SCI-REALITY-0388
Wait, what? A sensor has a range from 0 to 100 units. Its specification says accuracy: ±1% full scale. The display currently reads 10 units. A learner says, “One per cent of 10 is 0.1, so the reading must be accurate to ±0.1 unit.”
The arithmetic is correct for one per cent of the reading. But the specification did not say one per cent of the reading. It said one per cent of full scale. If full scale is 100 units, one per cent of full scale is 1 unit. The reference quantity has changed, so the meaning of the percentage has changed.
This Reality Lab owns one narrow learner job: how to read a percentage-based instrument specification by finding what the percentage is a percentage of. It does not re-teach measurement accuracy, resolution, calibration or uncertainty as general concepts. It applies those owners to a real-world datasheet phrase that can make a small number look more precise than the evidence supports.
Quick Answer
No. “±1% full scale” normally refers to one per cent of the instrument’s stated full-scale range for the relevant range or mode, not one per cent of each displayed reading. For a 0–100 unit range, ±1% full scale corresponds to ±1 unit under the stated specification conditions. At a reading of 10 units, that ±1 unit equals ten per cent of the reading. At a reading of 90 units, the same ±1 unit is only about 1.1 per cent of the reading.
Real instrument specifications can contain extra terms such as “±1 digit”, offsets, temperature effects or different ranges. Never simplify a real specification by deleting its other terms. The core habit here is simpler: find the denominator before interpreting a percentage.
The Datasheet Object
Create an original fictional sensor card:
| Quantity | Pressure |
|---|---|
| Range | 0–100 kPa |
| Resolution | 0.1 kPa |
| Accuracy specification | ±1% full scale |
| Operating condition | Specified laboratory temperature range |
Now ask three different questions:
- What is full scale? 100 kPa.
- What is one per cent of full scale? 1 kPa.
- Does a display of 10.0 kPa make the accuracy term 0.1 kPa? Not from this specification.
The percentage belongs to the range reference, not automatically to the current displayed value.
A Real-World Clue From NIST Instrument Pages
NIST instrument pages provide a useful reality check because they show that manufacturers and laboratories can specify performance in different ways. One NIST-listed conductivity meter reports an accuracy term as a percentage of full scale plus a digit term, while another NIST-listed meter reports a percentage of the measurement value. A third example uses full scale for one measurement mode and percentage of reading for another.
That variation is exactly why students should not see “1%” and stop reading. The words after the percentage carry scientific meaning.
Observed, Specified and Inferred
- Observed: the instrument display shows a value, for example 10.0 kPa.
- Specified: the datasheet gives an accuracy expression, for example ±1% full scale, under stated conditions.
- Calculated: on a 0–100 kPa range, 1% full scale equals 1 kPa.
- Invalid inference: “Because the display is 10, the error term is 1% of 10.”
A percentage without its reference quantity is incomplete evidence.
Why the Same Specification Feels Bigger Near the Bottom of the Range
Suppose the full-scale term is fixed at ±1 unit on a 0–100 unit range. Compare three readings:
| Displayed reading | ±1% full scale term | Size of that term relative to reading |
|---|---|---|
| 10 units | ±1 unit | 10% of the reading |
| 50 units | ±1 unit | 2% of the reading |
| 90 units | ±1 unit | about 1.1% of the reading |
This does not mean the instrument suddenly becomes dishonest at low readings. It means the stated full-scale term is referenced to the same range size while the measured value changes. NIST measurement guidance has long warned that full-scale-based performance can become a large fraction of a small measured quantity, which is one reason range selection matters.
Worked Case 1: Two Sensors Both Say “1%”
Sensor A: range 0–100 units, accuracy ±1% full scale. Sensor B: range 0–100 units, accuracy ±1% of reading. Both display 20 units.
For Sensor A, the stated percentage term is ±1 unit. For Sensor B, the percentage term is ±0.2 unit. Therefore “both are 1% accurate” hides a crucial difference.
Better comparison: translate both specifications into the same absolute unit at the reading of interest, while keeping any additional terms and stated conditions.
Worked Case 2: Same Sensor, Different Range
A meter can operate on either a 0–10 V range or a 0–100 V range. Imagine the specification is ±1% of the selected full scale. You want to measure about 8 V.
On the 0–10 V range, one per cent full scale is 0.1 V. On the 0–100 V range, one per cent full scale is 1 V. The same 8 V signal can therefore have a much smaller stated full-scale term when the appropriate lower range is used, assuming the lower range safely includes the measurement and all other specification conditions are satisfied.
This is why instrument range is part of the evidence, not merely a display preference.
Worked Case 3: Resolution Is Not the Same as the Accuracy Term
A sensor displays values to 0.01 unit but has an accuracy specification of ±1 unit. A learner says, “It shows two decimal places, so the true value must be known to the nearest 0.01.”
The display resolution tells you the smallest displayed increment. It does not prove the measurement is accurate to that increment. A display can show 10.37 while the stated performance allows a much wider difference from a suitable reference.
Worked Case 4: The Extra “±1 Digit”
Some real specifications combine terms, for example a percentage term plus one or more digits. A student copies only “±2% full scale” and ignores the digit term because the percentage looks more important.
That changes the specification. When a manufacturer or laboratory gives a combined expression, the scientific reader must preserve the entire expression and its stated conditions. Reality Lab does not need advanced error propagation to teach this habit: do not delete inconvenient parts of the evidence.
Worked Case 5: A Low Reading Is Not Automatically Useless
Suppose a 0–100 unit instrument with ±1 unit full-scale term reads 5 units. The ±1 unit term is large relative to 5. A learner says, “Then the measurement is worthless.”
That conclusion is also too strong. Whether the measurement is useful depends on the decision. If you only need to know whether the quantity is roughly near 5 or near 80, it may be entirely useful. If you need to distinguish 5.0 from 5.2, the instrument may be unsuitable. Evidence quality is always matched to the job.
Worked Case 6: An Advertisement Removes the Denominator
A product comparison says, “Model X: 1% accuracy. Model Y: 2% accuracy.” The original datasheets reveal that X is ±1% full scale on a 0–100 range, while Y is ±2% of reading over the relevant range.
The advertisement has made unlike expressions look directly comparable by hiding what each percentage refers to. To evaluate the claim, choose a realistic reading, preserve all terms, calculate both specifications in the same units, and compare under matched conditions.
The Baseline Check: Percentage of What?
Students often treat a percentage as self-explanatory. In science, a percentage is a ratio and therefore needs a denominator or reference amount. “Ten per cent” could mean ten per cent of:
- the current reading;
- full-scale range;
- a reference value;
- an initial value;
- a control result;
- the total sample;
- a theoretical maximum.
Change the denominator and the scientific meaning changes.
Method Check
Before turning a datasheet accuracy line into a claim about one result, check:
- What range was selected?
- What exactly does “full scale” mean for that range?
- Is the percentage of full scale, of reading, or a combination?
- Are there added constant, digit or offset terms?
- What temperature or environmental conditions apply?
- Does the specification refer to a sensor alone or the whole measurement system?
- Has the instrument been calibrated and used within its stated conditions?
- Is the question about specification, actual uncertainty of one result, repeatability or resolution?
Representation Check: Why “±1%” Looks Smaller Than It Can Be
A marketing table may shorten “±1% full scale” to a large bold “1%”. That makes the number visually simple while removing the reference. The full phrase is the evidence object. If the denominator disappears, the reader can accidentally substitute the denominator that feels natural.
The repair is mechanical: rewrite the specification as a sentence. “The stated percentage term equals one per cent of the selected full-scale range.” Then calculate it in physical units.
Alternative Explanations When a Reading Disagrees With a Reference
Even after interpreting full-scale percentage correctly, a disagreement can have several causes:
- the instrument is operating near the edge of its stated conditions;
- the wrong range was selected;
- the reference instrument has its own uncertainty;
- the measured quantity changed between readings;
- the sensor has drifted since calibration;
- installation or sample handling changed the measurand;
- the specification includes additional terms that were omitted;
- the reading is being compared with a reference that represents a different place or time.
Do not force every disagreement into a single explanation.
What Evidence Strengthens a Datasheet-Based Claim?
- the full range and selected range are stated;
- the accuracy expression is copied completely;
- units are converted consistently;
- operating conditions match the datasheet conditions;
- the instrument has a relevant calibration or reference check;
- the measurement is comfortably inside the instrument’s suitable range;
- independent checks agree within expected limits;
- the claim is limited to what the specification actually addresses.
What Weakens an Overclaim?
- “1%” is shown with no denominator;
- full scale is mistaken for the current reading;
- decimal places are treated as proof of accuracy;
- additional specification terms are deleted;
- a specification is treated as the exact uncertainty of every individual reading;
- conditions such as temperature or range are ignored;
- two instruments are compared using unlike accuracy definitions;
- a low-range measurement is made on a needlessly wide range and then called “precise”.
How Far Can the Conclusion Travel?
From a datasheet line “±1% full scale” and a verified 0–100 unit range, you can calculate the stated percentage term as ±1 unit. You cannot automatically claim that the true value is guaranteed inside that interval under every condition, that the actual uncertainty of one reading is exactly ±1 unit, that the instrument’s resolution is 1 unit, or that another range has the same absolute term.
The safe conclusion is bounded: “On this selected 0–100 unit range, the stated ±1% full-scale term corresponds to ±1 unit under the manufacturer’s specified conditions. Other terms and real measurement conditions still matter.”
Tempting but Invalid Reasoning
- “1% is always 1% of the reading.” The specification must state its reference.
- “A reading of 10 with ±1% FS means ±0.1.” Only if full scale were 10; on a 100-unit range it is ±1.
- “More decimal places mean a smaller accuracy error.” Display resolution and accuracy are different.
- “The instrument is 99% correct.” An accuracy specification is not a percentage score of correct readings.
- “The lowest range is always best.” The range must still safely contain the measurand and meet the instrument’s conditions.
- “The datasheet gives the exact error of this measurement.” Specifications bound or describe performance under stated conditions; the actual error of one result is not directly known merely from the spec.
PSLE-Style Transfer Case
A school uses two force sensors. Sensor P has a 0–50 N range and a stated term of ±2% full scale. Sensor Q has a 0–50 N range and a stated term of ±2% of reading. Both display 5 N.
For P, two per cent of full scale is 1 N. For Q, two per cent of the reading is 0.1 N. Therefore the equal-looking “2%” labels do not represent equal absolute terms at 5 N. A student should not decide which sensor is better overall from this alone; range, resolution, additional terms, calibration, environment and the required measurement job also matter.
Delayed Independent Return
- A 0–200 unit instrument is specified at ±0.5% full scale. What is the percentage term in units?
- If it reads 20 units, is the term 0.5% of 20?
- Why can full-scale accuracy look relatively large at low readings?
- Why is resolution not the same as accuracy?
- What words must remain attached to a percentage before you interpret it?
- What should you do before comparing “1% accuracy” from two different instruments?
Explained Answers
1. One unit, because 0.5% of 200 is 1. 2. No; not when the specification says full scale. 3. The absolute full-scale term can stay fixed while the reading becomes smaller. 4. Resolution tells how finely the instrument displays or distinguishes values; accuracy concerns agreement with a reference under stated conditions. 5. The reference quantity: full scale, reading, range or other denominator, plus any additional terms. 6. Translate both complete specifications into comparable units at the same relevant reading and conditions.
Route the Core Skills to Their Owners
For the difference between an accuracy specification and the uncertainty of one result, use PSLE Science Reality Lab Vol No.151 | “Accuracy ±1°C” — Is That the Uncertainty of This One Reading?. For instrument range and resolution, use How to Choose a Measuring Instrument for PSLE Science That Has the Right Range and Resolution. For accuracy versus precision, use How to Tell Measurement Precision From Accuracy in PSLE Science Without Assuming Repeated Agreement Means Correct.
Parent and Tutor Teaching Guide
Write three cards: “1% of reading”, “1% full scale on 0–100”, and “±1 unit”. Ask the learner which two are equivalent. Then change the reading from 10 to 80 without changing the range. The full-scale term remains 1 unit while one per cent of reading changes from 0.1 to 0.8. That contrast exposes the denominator directly.
Next, give two fictional instruments with different ranges and ask the learner to convert each percentage specification into physical units before choosing one for a 7-unit measurement. Do not ask “Which instrument is better?” until the job is defined.
Finally, show a display with many decimal places and a much wider stated accuracy term. Ask which number tells display resolution and which describes stated performance. This prevents children from equating visual precision with scientific accuracy.
Authoritative Sources
- NIST — Horiba Twin Model B-173 Conductivity Meter, an example of a real instrument specification using a percentage of full scale plus a digit term.
- NIST — ATI Orion Model 130 Conductivity Meter, showing a real specification expressed as a percentage of measurement value, useful for contrasting denominators.
- NIST — EcoSense EC300A Conductivity Meter, showing that different modes can use different forms such as percent of reading or percent full scale.
- Singapore Examinations and Assessment Board — 2026 PSLE Science syllabus.
- Ministry of Education Singapore — 2023 Primary Science Teaching & Learning Syllabus.
Quiet Return
A percentage is never floating in space. It is a fraction of something. When a scientific instrument says “±1% full scale”, pause before multiplying the displayed reading by 0.01. Find the scale, preserve every term, convert the specification into the same physical units as the measurement, and only then decide what the evidence supports. The denominator is part of the science.
