PSLE Science Reality Lab Vol No.250 | “Air Sensor Says PM2.5 = 40” — Is That the Same as a Regulatory Monitor Reading?

Stable internal ID: PSLE-SCI-REALITY-0250

Wait, what? A small air sensor on a school balcony shows PM2.5 = 40 µg/m³. A larger official monitoring station several kilometres away reports a different value at about the same time. One pupil says, “The small sensor must be wrong.” Another says, “No, the official monitor must be wrong because the sensor is right here.”

Both reactions move too quickly. Before deciding which number deserves more weight, a scientist asks a more useful question: what kind of measurement system produced each number, under what conditions, for what purpose, and with what performance evidence?

The U.S. Environmental Protection Agency distinguishes air sensors used for non-regulatory supplemental and informational monitoring from the reference and equivalent methods used for regulatory monitoring. EPA also publishes voluntary performance-testing protocols for air sensors. That does not make a small sensor useless. It means the job of the instrument matters.

This is exactly the kind of evidence discipline that belongs in PSLE Science. The 2026 PSLE Science assessment objectives include interpreting and analysing information, evaluating observations, information and methods, and communicating explanations and reasoning. The habit is not “trust the expensive machine.” The habit is “match the claim to the measurement system.”

Quick Answer

Not automatically. A low-cost or compact air sensor can provide useful measurements, trends and local information, but its reading is not automatically equivalent to a regulatory reference-monitor result.

Check the pollutant, units, location, averaging time, sensor principle, environmental conditions, maintenance, calibration or correction, performance testing and whether the sensor has been compared beside a suitable reference instrument. If two devices disagree, do not immediately assume one is defective. They may be measuring different air, different time windows or the same air with different methods and uncertainties.

The Owned Learner Job

This Reality Lab owns one narrow real-world evidence-transfer job: how to evaluate an air-sensor display, comparison chart or product claim without treating every instrument that prints the same unit as an interchangeable measurement system.

It does not own the science of particle formation, respiratory health, air-pollution regulation, sensor engineering or PM2.5 as a standalone concept. It also does not tell you whether outdoor air is safe for a particular person. Those questions belong to relevant scientific, medical and public-health authorities. Here, the air sensor is a communication object through which we practise evidence reasoning.

Rebuild the Evidence Object: Two Numbers That Look Identical

Imagine an original composite case. Sensor A is a compact optical particle sensor mounted outside a classroom. It reports PM2.5 every minute. Monitor B is an official station using a method designed for a formal monitoring network. At 3:00 p.m., Sensor A displays 40 µg/m³. The network page reports 28 µg/m³.

The tempting question is, “Which number is correct?” A better first set of questions is:

Are the two values for the same exact minute or different averaging periods?
Are the devices at the same location and height?
Do they measure particles using the same physical method?
Has Sensor A been tested against a suitable reference instrument?
Was a correction algorithm used?
Could humidity, temperature or particle type affect Sensor A differently?
Was either instrument undergoing maintenance or quality checks?

A scientific comparison begins by making the evidence comparable.

Observed, Reported and Inferred

Observed by the device: a physical response inside the sensor, such as light scattered by particles passing through an optical chamber.
Reported by the system: a processed concentration estimate in µg/m³ after the device applies its calibration or algorithm.
Inferred by the reader: statements such as “the air here is more polluted than everywhere nearby” or “this equals the official regulatory concentration.”

The first two can be legitimate steps in a measurement chain. The third can become too strong if we skip the method and comparison checks.

Same Unit Does Not Mean Same Measurement Quality

Two rulers can both report centimetres while differing in straightness, markings and calibration. Two thermometers can both display °C while responding differently to sunlight or contact. In the same way, two air instruments can both display µg/m³ without having identical measurement performance.

The unit tells you what quantity is being reported. It does not, by itself, tell you how accurately, precisely or robustly the quantity was measured.

What “Non-Regulatory Supplemental and Informational Monitoring” Means

EPA’s air-sensor guidance uses the phrase non-regulatory supplemental and informational monitoring for many sensor applications. Examples include observing trends, finding possible local hotspots, supporting citizen-science activities and adding spatial detail between larger monitoring stations.

The word non-regulatory does not mean “fake” or “scientifically worthless.” It means the sensor is being used for a different evidence job from a method used to make formal regulatory determinations. A useful scientific tool does not have to own every possible measurement job.

Performance Evidence Matters More Than the Price Tag

A cheap device is not automatically bad. An expensive device is not automatically perfect. Scientific confidence should come from evidence about performance.

Useful questions include:

Was the sensor tested in the field beside a suitable reference instrument?
Was it also tested under controlled laboratory conditions?
How well did it track changing concentrations?
Did it show a systematic high or low bias?
How did humidity and temperature affect it?
Did different units of the same model agree with one another?
Did performance drift over time?

EPA’s performance-testing reports were created precisely because sensor usefulness should be evaluated with consistent metrics and protocols rather than assumed from appearance or marketing.

Worked Case 1: The Humid Afternoon

Sensor A and a reference monitor agree reasonably well on dry mornings. On several very humid afternoons, Sensor A reads higher.

A weak conclusion is, “Air pollution always becomes much worse whenever humidity rises.” A stronger next step is to ask whether humidity changes the particles, the sensor’s optical response, or both. Some optical particle sensors can respond differently when particles take up water. The pattern is evidence, but the mechanism and correction need checking.

Notice the evidence habit: we do not erase the high reading, and we do not immediately turn it into a pollution event. We investigate why the measurement changed.

Worked Case 2: The Roadside Spike

A school sensor near a busy road shows a short spike while the official station several kilometres away does not.

It would be wrong to say, “The local sensor must be false because the official station did not see it.” The two instruments sampled different places. A short local event could affect one location without affecting the other.

But it would also be too strong to say, “The entire district had the same spike.” One local sensor cannot automatically describe a whole region. The conclusion can travel only as far as the sampling location and evidence justify.

Worked Case 3: The Correction Equation

A sensor originally reads 36 µg/m³. After scientists compare many readings with a reference instrument, they apply a correction and the adjusted value becomes 30 µg/m³.

Did someone “change the data to make it look nicer”? Not necessarily. A correction can be scientifically justified when a measurement system has a known systematic response and the correction is derived and validated appropriately.

The questions become: How was the correction obtained? Under what conditions? Does it still work at other seasons, particle mixtures or humidity levels? Was the raw reading preserved? Scientific processing should be traceable, not mysterious.

Worked Case 4: Same Model, Different Units

Three identical sensors are placed side by side. They report 24, 25 and 31 µg/m³. That disagreement is evidence.

Do not average immediately and forget the spread. First ask whether 31 is within the expected unit-to-unit variation, whether one sensor needs maintenance, whether air flow is obstructed, or whether one device has drifted. Replicate instruments can reveal uncertainty that a single display hides.

Worked Case 5: A Trend Can Be Useful Even When the Absolute Value Is Imperfect

Suppose a sensor reads slightly high compared with a reference instrument but responds consistently when local particle levels rise and fall. It may still be useful for detecting patterns or timing changes if that use has been evaluated.

This is an important scientific idea: a tool can be fit for one purpose without being fit for every purpose. “Useful” and “perfect” are not the only two choices.

Representation Check: What Does the Dashboard Actually Show?

Is the displayed number raw or corrected?
Is it a one-minute reading, an hourly average or a daily average?
Does the map interpolate between sensors?
Are missing values hidden?
Does a colour category represent concentration or an index?
Is the timestamp local and current?
Has the sensor been flagged for maintenance?

A dashboard is not the atmosphere itself. It is a representation built from instruments, algorithms, time windows and display choices.

Location Check: Near Is Not the Same as Together

Air can vary over short distances near roads, construction, cooking exhaust, fires, sea spray or other sources. If two instruments are kilometres apart, disagreement may be partly spatial rather than purely instrumental.

For a direct instrument comparison, scientists often collocate devices—place them very close together so they sample nearly the same air. This reduces one major alternative explanation.

Time Check: One Minute Is Not One Day

A one-minute peak can disappear inside an hourly or daily average. Conversely, an hourly average cannot tell you the exact highest one-minute value.

Before comparing two numbers, align their averaging periods. This is the same scientific discipline used when comparing temperatures, river flows, growth rates or any changing quantity.

Method Check: What Physical Signal Is the Sensor Using?

Many compact particle sensors shine light through air and estimate particle concentration from scattered light. Reference methods may use different physical principles or carefully specified measurement procedures. Different methods can respond differently to particle size, shape, composition and moisture.

That does not mean one method is “lying.” It means measurement is a designed scientific process. Understanding the process helps explain when two numbers can be compared directly and when extra validation is needed.

What Evidence Strengthens an Air-Sensor Claim?

Field collocation with a suitable reference instrument.
Testing across a realistic concentration range.
Testing across humidity and temperature conditions relevant to use.
Repeat testing with more than one unit of the sensor model.
Transparent correction methods.
Quality checks showing the sensor remains stable over time.
Clear location, time and averaging information.
A conclusion limited to the sensor’s demonstrated use.

What Weakens It?

A single impressive reading with no comparison evidence.
A claim that the sensor is “official-grade” without a defined performance basis.
Comparing devices at different places and blaming all disagreement on accuracy.
Ignoring humidity, temperature or maintenance.
Changing correction formulas without documenting them.
Using a sensor validated for one pollutant or environment to make claims about another.
Turning a measurement into medical or legal advice it was not designed to provide.

How Far Can the Conclusion Travel?

A careful conclusion might be:

This sensor recorded a higher PM2.5 estimate at this location and time. The reading should be interpreted using the device’s performance evidence, environmental conditions and averaging period, and it is not automatically interchangeable with a regulatory reference-monitor measurement.

Do not automatically upgrade that to “the whole city had the same concentration,” “the regulatory station is wrong,” “the sensor proves a health effect,” or “the sensor is approved for every monitoring purpose.”

Tempting but Invalid Reasoning

“It has decimals, so it must be highly accurate.” Display resolution is not the same as accuracy.
“The official monitor is farther away, so it is less scientific.” Distance and measurement quality are different questions.
“The cheap sensor disagrees, so it is useless.” A tool can still be useful for trends or local variation when its performance is understood.
“Both say µg/m³, so the numbers are directly interchangeable.” Same units do not erase method differences.
“EPA has a sensor protocol, so a tested sensor is EPA certified.” EPA states that its voluntary testing protocols do not themselves constitute certification or endorsement.

PSLE-Style Transfer Case: Three Sensors Around a School

Three identical particle sensors are placed at the front gate, courtyard and rooftop. At 8:00 a.m. they report 42, 29 and 25 µg/m³. A pupil concludes, “The front sensor is inaccurate because it is different from the other two.”

A stronger answer is:

The front sensor may be measuring genuinely different air, for example because it is nearer a local particle source. To test whether the difference is due to the sensor itself, place the sensors together under the same conditions and compare their readings with one another and, where appropriate, with a suitable reference instrument.

This answer protects two possibilities at once: spatial variation and instrument variation.

Explained Practice

Practice 1

A sensor matches a reference monitor on dry days but reads higher on humid days. What should you investigate before claiming pollution always increased?

Answer: Check whether humidity affects the sensor response, whether the particle mixture changed, and whether the instruments were collocated over the same averaging period.

Practice 2

A map combines one-minute sensor readings with hourly official-station averages. Can you compare the numbers directly?

Answer: Not without aligning the time windows. A short peak and an hourly mean describe different summaries.

Practice 3

A manufacturer says its sensor met voluntary EPA performance targets. Does that mean EPA endorses the product?

Answer: No. EPA states that the testing recommendations are voluntary and results do not constitute EPA certification or endorsement.

Practice 4

Two sensors disagree by 3 µg/m³ while side by side. Is that automatically a serious failure?

Answer: No. Compare the difference with the expected performance, concentration range and uncertainty. The size of a disagreement matters in context.

Delayed Independent Return: The Three-Layer Reading

Tomorrow, take any environmental sensor display and write three lines without notes: physical signal → reported quantity → claim being made. Then mark where calibration, correction or inference enters the chain. If you can do that, you are no longer treating a digital number as if it arrived directly from nature with no measurement system in between.

Parent and Tutor Teaching Guide

You do not need an air-quality sensor to teach this lesson. Put two imaginary instruments on paper. Give them the same unit but different locations, time windows and methods. Ask the learner what must be made comparable before deciding which reading is stronger evidence.

Then remove one difference at a time. Place the devices together. Match the averaging period. Add a reference instrument. Add repeated observations. The learner sees how experimental control strengthens a comparison without being handed a memorised slogan.

The desired habit is calm: a sensor reading is evidence produced by a measurement system. Evaluate the system before expanding the claim.

Route to Existing Canonical PSLE Science Owners

Authoritative Sources

Quiet Return

A small sensor can reveal something important. An official monitor can provide a different kind of evidence. The scientific move is not to choose a favourite machine before looking at the method.

When two numbers disagree, ask what each number actually represents, how it was produced, and what claim the measurement system was designed to support.

Small Group Tutorials